Package 'datarobot'

Description

For working with the 'DataRobot' predictive modeling platform's API https://www.datarobot.com/.

Description

Each Eureqa model contains multiple possible solutions (see GetParetoFront). However, only the best model is included in the leaderboard by default. To include other models, you can get them via GetParetoFront and then add them.

Usage

AddEureqaSolution(project, eureqaSolutionId)

Arguments

Examples

## Not run: 
  projectId <- "5b2827556523cd05bd1507a5"
  modelId <- "5b29406c6523cd0665685a8d"
  eureqaModel <- GetModel(projectId, modelId)
  paretoFront <- GetParetoFront(eureqaModel)

## End(Not run)

Description

Apply a schema to DataRobot objects (lists, frames)

Usage

ApplySchema(inList, schema, mask = NULL)

Arguments

Description

These functions extend R's generic as.data.frame function to the DataRobot S3 object classes listOfBlueprints, listOfFeaturelists, listOfModels, and projectSummaryList.

If simple = TRUE (the default), this method returns a dataframe with one row for each model and the following columns: projectName, projectId, created, fileName, target, targetType, positiveClass, metric, autopilotMode, stage, maxTrainPct, and holdoutUnlocked. If simple = FALSE, a dataframe is constructed from all elements of projectSummaryList.

Usage

## S3 method for class 'listOfBlueprints'
as.data.frame(x, row.names = NULL, optional = FALSE, ...)

## S3 method for class 'listOfFeaturelists'
as.data.frame(x, row.names = NULL, optional = FALSE, ...)

## S3 method for class 'listOfModels'
as.data.frame(x, row.names = NULL, optional = FALSE, simple = TRUE, ...)

## S3 method for class 'projectSummaryList'
as.data.frame(x, row.names = NULL, optional = FALSE, simple = TRUE, ...)

## S3 method for class 'listOfDataRobotPredictionDatasets'
as.data.frame(x, row.names = NULL, optional = FALSE, ...)

Arguments

Details

All of the DataRobot S3 ‘listOf’ class objects have relatively complex structures and are often easier to work with as dataframes. The methods described here extend R's generic as.data.frame function to convert objects of these classes to convenient dataframes. For objects of class listOfBlueprints and listOfFeaturelists or objects of class listOfModels and projectSummaryList with simple = FALSE, the dataframes contain all information from the original S3 object. The default value simple = TRUE provides simpler dataframes for objects of class listOfModels and projectSummaryList.

If simple = TRUE (the default), this method returns a dataframe with one row for each model and the following columns: modelType, expandedModel (constructed from modelType and processes from the listOfModels elements), modelId, blueprintId, featurelistName, featurelistId, samplePct, and the metrics validation value for projectMetric. If simple = FALSE, the method returns a complete dataframe with one row for each model and columns constructed from all fields in the original listOfModels object

Value

A dataframe containing some or all of the data from the original S3 object; see Details.

Description

Information on a data feature.

Usage

as.dataRobotFeatureInfo(inList)

Arguments

Value

A named list which contains:

• id numeric. feature id. Note that throughout the API, features are specified using their names, not this ID.

• name character. The name of the feature.

• featureType character. Feature type: 'Numeric', 'Categorical', etc.

• importance numeric. numeric measure of the strength of relationship between the feature and target (independent of any model or other features).

• lowInformation logical. Whether the feature has too few values to be informative.

• uniqueCount numeric. The number of unique values in the feature.

• naCount numeric. The number of missing values in the feature.

• dateFormat character. The format of the feature if it is date-time feature.

• projectId character. Character id of the project the feature belongs to.

• max. The maximum value in the dataset, formatted in the same format as the data.

• min. The minimum value in the dataset, formatted in the same format as the data.

• mean. The arithmetic mean of the dataset, formatted in the same format as the data.

• median. The median of the dataset, formatted in the same format as the data.

• stdDev. The standard deviation of the dataset, formatted in the same format as the data.

• timeSeriesEligible logical. Whether this feature can be used as the datetime partition column in a time series project.

• timeSeriesEligibilityReason character. Why the feature is ineligible for the datetime partition column in a time series project, "suitable" when it is eligible.

• crossSeriesEligible logical. Whether the cross series group by column is eligible for cross-series modeling. Will be NULL if no cross series group by column is used.

• crossSeriesEligibilityReason character. The type of cross series eligibility (or ineligibility).

• timeStep numeric. For time-series eligible features, a positive integer determining the interval at which windows can be specified. If used as the datetime partition column on a time series project, the feature derivation and forecast windows must start and end at an integer multiple of this value. NULL for features that are not time series eligible.

• timeUnit character. For time series eligible features, the time unit covered by a single time step, e.g. "HOUR", or NULL for features that are not time series eligible.

• targetLeakage character. Whether a feature is considered to have target leakage or not. A value of "SKIPPED_DETECTION" indicates that target leakage detection was not run on the feature.

• keySummary data.frame. Optional. Descriptive statistics for this feature, iff it is a summarized categorical feature. This data.frame contains:

  • key. The name of the key.

  • summary. Descriptive statistics for this key, including:

    • max. The maximum value in the dataset.

    • min. The minimum value in the dataset.

    • mean. The arithmetic mean of the dataset.

    • median. The median of the dataset.

    • stdDev. The standard deviation of the dataset.

    • pctRows. The percentage of rows (from the EDA sample) in which this key occurs.

See Also

Other feature functions: GetFeatureInfo(), ListFeatureInfo(), ListModelFeatures()

Description

Return value for GetMultiSeriesProperties() and others

Usage

as.dataRobotMultiSeriesProperties(inList)

Arguments

Value

A named list which contains:

• timeSeriesEligible logical. Whether or not the series is eligible to be used for time series.

• crossSeriesEligible logical. Whether or not the cross series group by column is eligible for cross-series modeling. Will be NULL if no cross series group by column is used.

• crossSeriesEligibilityReason character. The type of cross series eligibility (or ineligibility).

• timeUnit character. For time series eligible features, the time unit covered by a single time step, e.g. "HOUR", or NULL for features that are not time series eligible.

• timeStep integer. Expected difference in time units between rows in the data. Will be NULL for features that are not time series eligible.

See Also

Other MultiSeriesProject functions: GetMultiSeriesProperties(), RequestCrossSeriesDetection(), RequestMultiSeriesDetection()

Description

Return value for SetupProject() and others

Usage

as.dataRobotProjectShort(inList)

Arguments

Value

A named list that contains:

projectName

character. The name assigned to the DataRobot project

projectId

character. The unique alphanumeric project identifier for this DataRobot project

fileName

character. The name of the CSV modeling file uploaded for this project

created

character. The time and date of project creation

Description

This is a list that contains the valid values for autopilot mode. If you wish, you can specify autopilot modes using the list values, e.g. AutopilotMode$FullAuto instead of typing the string "auto". This way you can benefit from autocomplete and not have to remember the valid options.

Usage

AutopilotMode

Format

An object of class list of length 4.

Details

FullAuto represents running the entire autopilot. Quick runs a quicker, abridged version of the autopilot that focuses on the most important models. Manual does not run the autopilot and instead leaves it to the user to select the algorithms to be run. Comprehensive runs all blueprints in the repository, and may be extremely slow.

Description

Supports feature transformations, including:

• text to categorical

• text to numeric

• categorical to text

• categorical to numeric

• numeric to categorical

Usage

BatchFeaturesTypeTransform(
  project,
  parentNames,
  variableType,
  prefix = NULL,
  suffix = NULL,
  maxWait = 600
)

Arguments

Value

a list of all the features, after transformation. See GetFeaturelist for details.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  BatchFeaturesTypeTransform(projectId,
                             parentNames = c("var1", "var2"),
                             variableType = VariableTransformTypes$Categorical,
                             suffix = "_transformed")

## End(Not run)

Description

This is a list that contains the valid values for Blend methods

Usage

BlendMethods

Format

An object of class list of length 13.

Description

Convert a blueprint chart into graphviz DOT format

Usage

BlueprintChartToGraphviz(blueprintChart)

Arguments

Value

Character string representation of chart in graphviz DOT language.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  model <- GetModel(projectId, modelId)
  blueprintId <- model$blueprintId
  blueprintChart <- GetBlueprintChart(projectId, blueprintId)
  BlueprintChartToGraphviz(blueprintChart)

## End(Not run)

Description

Make sure the path is a reasonable URL

Usage

CheckUrl(url)

Arguments

Description

Added in DataRobot API 2.18.

Usage

ClassificationDeploymentAccuracyMetric

Format

An object of class list of length 14.

Description

Reformat paginated data returned from the server.

Usage

CleanServerData(serverData)

Arguments

Description

This function clones a project, creating a fresh (post-EDA1) copy that will need a target and modeling options set.

Usage

CloneProject(project, newProjectName = NULL, maxWait = 600)

Arguments

Value

A named list that contains:

projectName

character. The name assigned to the DataRobot project

projectId

character. The unique alphanumeric project identifier for this DataRobot project

fileName

character. The name of the CSV modeling file uploaded for this project

created

character. The time and date of project creation

Examples

## Not run: 
 project <- GetProject("5c1303269300d900016b41a7")
 CloneProject(project, newProjectName = "Project Restart")

## End(Not run)

Description

Compute datetime trend plots for datetime partitioned model. This includes Accuracy over Time, Forecast vs Actual, and Anomaly over Time plots.

Usage

ComputeDatetimeTrendPlots(
  model,
  backtest = 0,
  source = SourceType$Validation,
  forecastDistanceStart = NULL,
  forecastDistanceEnd = NULL
)

Arguments

Details

• Forecast distance specifies the number of time steps between the predicted point and the origin point.

• For the multiseries models only first 1000 series in alphabetical order and an average plot for them will be computed.

• Maximum 100 forecast distances can be requested for calculation in time series supervised projects.

Value

An integer value that can be used as the jobId parameter in a subsequent call to WaitForJobToComplete.

Examples

## Not run: 
projectId <- "59a5af20c80891534e3c2bde"
modelId <- "5996f820af07fc605e81ead4"
model <- GetModel(projectId, modelId)
jobId <- ComputeDatetimeTrendPlots(model)
WaitForJobToComplete(projectId, jobId) # optional step

## End(Not run)

Description

This function initializes a DataRobot session. To use DataRobot, you must connect to your account. This can be done in three ways:

• by passing an endpoint and token directly to ConnectToDataRobot

• by having a YAML config file in $HOME/.config/datarobot/drconfig.yaml

• by setting DATAROBOT_API_ENDPOINT and DATAROBOT_API_TOKEN environment variables

The three methods of authentication are given priority in that order (explicitly passing parameters to the function will trump a YAML config file, which will trump the environment variables.) If you have a YAML config file or environment variables set, you will not need to pass any parameters to ConnectToDataRobot in order to connect.

Usage

ConnectToDataRobot(
  endpoint = NULL,
  token = NULL,
  username = NULL,
  password = NULL,
  userAgentSuffix = NULL,
  sslVerify = TRUE,
  configPath = NULL
)

Arguments

Examples

## Not run: 
  ConnectToDataRobot("https://app.datarobot.com/api/v2", "thisismyfaketoken")
  ConnectToDataRobot(configPath = "~/.config/datarobot/drconfig.yaml")

## End(Not run)

Description

A duration of six months, 3 days, and 12 hours could be represented as P6M3DT12H.

Usage

ConstructDurationString(
  years = 0,
  months = 0,
  days = 0,
  hours = 0,
  minutes = 0,
  seconds = 0
)

Arguments

Value

The duration string, specified compatibly with ISO8601.

Examples

ConstructDurationString()
ConstructDurationString(days = 100)
ConstructDurationString(years = 10, months = 2, days = 5, seconds = 12)

Description

Uniquely defines a Backtest used in a DatetimePartitioning

Usage

CreateBacktestSpecification(
  index,
  gapDuration,
  validationStartDate,
  validationDuration
)

Arguments

Details

Includes only the attributes of a backtest directly controllable by users. The other attributes are assigned by the DataRobot application based on the project dataset and the user-controlled settings. All durations should be specified with a duration string such as those returned by the ConstructDurationString helper function.

Value

list with backtest parameters

Examples

zeroDayDuration <- ConstructDurationString()
hundredDayDuration <- ConstructDurationString(days = 100)
CreateBacktestSpecification(index = 0,
                            gapDuration = zeroDayDuration,
                            validationStartDate = "1989-12-01",
                            validationDuration = hundredDayDuration)

Description

Create a calendar from an uploaded CSV.

Usage

CreateCalendar(
  dataSource,
  name = NULL,
  multiSeriesIdColumn = NULL,
  maxWait = 600
)

Arguments

Value

An S3 object of class "dataRobotCalendar"

Examples

## Not run: 
   CreateCalendar("inst/extdata/calendar.csv", name = "intlHolidayCalendar")

## End(Not run)
## Not run: 
   holidayCalendarDF <- as.data.frame(myCalendar)
   CreateCalendar(holidayCalendarDF, name = "intlHolidayCalendar")

## End(Not run)
## Not run: 
   CreateCalendar("inst/extdata/calendar.csv",
                  name = "intlHolidayCalendar",
                  multiSeriesIdColumn = "Country")

## End(Not run)

Description

Note that if you're looking to download compliance documentation to a DOCX file, you can call DownloadComplianceDocumentation directly without using this function.

Usage

CreateComplianceDocumentation(model, templateId = NULL)

Arguments

Value

An integer value that can be used as the jobId parameter in a subsequent call to WaitForJobToComplete.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  model <- GetModel(projectId, modelId)
  jobId <- CreateComplianceDocumentation(model) # optional step
  WaitForJobToComplete(projectId, jobId)        # optional step
  DownloadComplianceDocumentation(model)

## End(Not run)

Description

Create a data source.

Usage

CreateDataSource(
  type,
  canonicalName,
  dataStoreId,
  query = NULL,
  table = NULL,
  schema = NULL,
  partitionColumn = NULL,
  fetchSize = NULL
)

Arguments

Examples

## Not run: 
 dataStoreId <- "5c1303269300d900016b41a7"
 CreateDataSource(type = "jdbc",
                  canonicalName = "Airline stats after 1995",
                  dataStoreId = dataStoreId,
                  query = 'SELECT * FROM airlines10mb WHERE "Year" >= 1995;')

## End(Not run)

Description

Create a data store.

Usage

CreateDataStore(type, canonicalName, driverId, jdbcUrl)

Arguments

Examples

## Not run: 
 CreateDataStore(type = "jdbc",
                 canonicalName = "Demo DB",
                 driverId = "57a7c978c808916f4a630f89",
                 jdbcUrl = "jdbc:postgresql://my.db.address.org:5432/my_db")

## End(Not run)

Description

Uniquely defines a DatetimePartitioning for some project

Usage

CreateDatetimePartitionSpecification(
  datetimePartitionColumn,
  autopilotDataSelectionMethod = NULL,
  validationDuration = NULL,
  holdoutStartDate = NULL,
  holdoutDuration = NULL,
  disableHoldout = NULL,
  gapDuration = NULL,
  numberOfBacktests = NULL,
  backtests = NULL,
  useTimeSeries = FALSE,
  defaultToKnownInAdvance = FALSE,
  featureDerivationWindowStart = NULL,
  featureDerivationWindowEnd = NULL,
  featureSettings = NULL,
  treatAsExponential = NULL,
  differencingMethod = NULL,
  windowsBasisUnit = NULL,
  periodicities = NULL,
  forecastWindowStart = NULL,
  forecastWindowEnd = NULL,
  multiseriesIdColumns = NULL,
  useCrossSeries = NULL,
  aggregationType = NULL,
  crossSeriesGroupByColumns = NULL,
  calendar = NULL
)

Arguments

Details

Includes only the attributes of DatetimePartitioning that are directly controllable by users, not those determined by the DataRobot application based on the project dataset and the user-controlled settings. This is the specification that should be passed to SetTarget via the partition parameter. To see the full partitioning based on the project dataset, GenerateDatetimePartition. All durations should be specified with a duration string such as those returned by the ConstructDurationString helper function.

Value

An S3 object of class 'partition' including the parameters required by the SetTarget function to generate a datetime partitioning of the modeling dataset.

Examples

CreateDatetimePartitionSpecification("date_col")
CreateDatetimePartitionSpecification("date",
                                     featureSettings = list(
                                       list("featureName" = "Product_offers",
                                            "defaultToKnownInAdvance" = TRUE)))
partition <- CreateDatetimePartitionSpecification("dateColumn",
                                                treatAsExponential = TreatAsExponential$Always,
                                                differencingMethod = DifferencingMethod$Seasonal,
                                                periodicities = list(list("timeSteps" = 10,
                                                                          "timeUnit" = "HOUR"),
                                                                     list("timeSteps" = 600,
                                                                          "timeUnit" = "MINUTE"),
                                                                     list("timeSteps" = 7,
                                                                          "timeUnit" = "DAY")))

Description

Create a deployment.

Usage

CreateDeployment(
  model,
  label = "",
  description = "",
  defaultPredictionServerId = NULL
)

Arguments

Value

A DataRobotDeployment object containing:

• id character. The ID of the deployment.

• label character. The label of the deployment.

• description character. The description of the deployment.

• defaultPredictionServer list. Information on the default prediction server connected with the deployment. See ListPredictionServers for details.

• model dataRobotModel. The model associated with the deployment. See GetModel for details.

• capabilities list. Information on the capabilities of the deployment.

• predictionUsage list. Information on the prediction usage of the deployment.

• permissions list. User's permissions on the deployment.

• serviceHealth list. Information on the service health of the deployment.

• modelHealth list. Information on the model health of the deployment.

• accuracyHealth list. Information on the accuracy health of the deployment.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  model <- GetModel(projectId, modelId)
  predictionServer <- ListPredictionServers()[[1]]
  CreateDeployment(model,
                   label = "myDeployment",
                   description = "this is my deployment",
                   defaultPredictionServerId = predictionServer)

## End(Not run)

Description

These functions request that new features be created as transformations of existing features and wait for the new feature to be created.

Usage

CreateDerivedFeatureAsCategorical(
  project,
  parentName,
  name = NULL,
  dateExtraction = NULL,
  replacement = NULL,
  maxWait = 600
)

CreateDerivedFeatureAsText(
  project,
  parentName,
  name = NULL,
  dateExtraction = NULL,
  replacement = NULL,
  maxWait = 600
)

CreateDerivedFeatureAsNumeric(
  project,
  parentName,
  name = NULL,
  dateExtraction = NULL,
  replacement = NULL,
  maxWait = 600
)

CreateDerivedFeatureIntAsCategorical(
  project,
  parentName,
  name = NULL,
  dateExtraction = NULL,
  replacement = NULL,
  maxWait = 600
)

Arguments

Value

Details for the created feature; same schema as the object returned from GetFeatureInfo.

Description

This function allows the user to create a new featurelist in a project by specifying its name and a list of variables to be included

Usage

CreateFeaturelist(project, listName, featureNames)

Arguments

Details

DataRobot featurelists define the variables from the modeling dataset used in fitting each project model. Some functions (SetTarget, StartNewAutopilot) optionally accept a featurelist (and use a default featurelist if none is specified).

Value

A list with the following four elements describing the featurelist created:

featurelistId

Character string giving the unique alphanumeric identifier for the new featurelist.

projectId

Character string giving the projectId identifying the project to which the featurelist was added.

features

Character vector with the names of the variables included in the new featurelist.

name

Character string giving the name of the new featurelist.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  CreateFeaturelist(projectId, "myFeaturelist", c("feature1", "feature2", "otherFeature"))

## End(Not run)

Description

Group partitioning constructs data partitions such that all records with each level in the column specified by the parameter partitionKeyCols occur together in the same partition.

Usage

CreateGroupPartition(
  validationType,
  holdoutPct,
  partitionKeyCols,
  reps = NULL,
  validationPct = NULL
)

Arguments

Details

This function is one of several convenience functions provided to simplify the task of starting modeling projects with custom partitioning options. The other functions are CreateRandomPartition, CreateStratifiedPartition, and CreateUserPartition.

Value

An S3 object of class 'partition' including the parameters required by the SetTarget function to generate a group-based partitioning of the modeling dataset.

See Also

CreateRandomPartition, CreateStratifiedPartition, CreateUserPartition.

Examples

CreateGroupPartition(validationType = "CV",
                     holdoutPct = 20,
                     partitionKeyCols = list("groupId"),
                     reps = 5)

Description

In time series projects, a new set of modeling features is created after setting the partitioning options. These features are automatically derived from those in the project's dataset and are the features used for modeling. Modeling features are only accessible once the target and partitioning options have been set. In projects that don't use time series modeling, once the target has been set, ModelingFeaturelists and Featurelists will behave the same.

Usage

CreateModelingFeaturelist(project, listName, featureNames)

Arguments

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  CreateModelingFeaturelist(projectId, "myFeaturelist", c("feature1", "feature2"))

## End(Not run)

Description

Create and validate the downloadable code for the ruleset associated with this model

Usage

CreatePrimeCode(project, primeModelId, language)

Arguments

Value

job Id

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  CreatePrimeCode(projectId, modelId, "Python")

## End(Not run)

Description

Random partitioning is supported for either Training/Validation/Holdout ("TVH") or cross-validation ("CV") splits. In either case, the holdout percentage (holdoutPct) must be specified; for the "CV" method, the number of cross-validation folds (reps) must also be specified, while for the "TVH" method, the validation subset percentage (validationPct) must be specified.

Usage

CreateRandomPartition(
  validationType,
  holdoutPct,
  reps = NULL,
  validationPct = NULL
)

Arguments

Details

This function is one of several convenience functions provided to simplify the task of starting modeling projects with custom partitioning options. The other functions are CreateGroupPartition, CreateStratifiedPartition, and CreateUserPartition.

Value

An S3 object of class partition including the parameters required by SetTarget to generate a random partitioning of the modeling dataset.

See Also

CreateStratifiedPartition, CreateGroupPartition, CreateUserPartition.

Examples

CreateRandomPartition(validationType = "CV", holdoutPct = 20, reps = 5)

Description

Creates and validates a new rating table from an uploaded CSV.

Usage

CreateRatingTable(
  project,
  parentModelId,
  dataSource,
  ratingTableName = "Uploaded Rating Table"
)

Arguments

Value

An integer value that can be used as the JobId parameter in subsequent calls representing this job.

Examples

## Not run: 
   projectId <- "5984b4d7100d2b31c1166529"
   modelId <- "5984b4d7100d2b31c1166529"
   CreateRatingTable(projectId, modelId, dataSource = "myRatingTable.csv")

## End(Not run)

Description

Stratified partitioning is supported for binary classification problems and it randomly partitions the modeling data, keeping the percentage of positive class observations in each partition the same as in the original dataset. Stratified partitioning is supported for either Training/Validation/Holdout ("TVH") or cross-validation ("CV") splits. In either case, the holdout percentage (holdoutPct) must be specified; for the "CV" method, the number of cross-validation folds (reps) must also be specified, while for the "TVH" method, the validation subset percentage (validationPct) must be specified.

Usage

CreateStratifiedPartition(
  validationType,
  holdoutPct,
  reps = NULL,
  validationPct = NULL
)

Arguments

Details

This function is one of several convenience functions provided to simplify the task of starting modeling projects with custom partitioning options. The other functions are CreateGroupPartition, CreateRandomPartition, and CreateUserPartition.

Value

An S3 object of class 'partition' including the parameters required by the SetTarget function to generate a stratified partitioning of the modeling dataset.

See Also

CreateGroupPartition, CreateRandomPartition, CreateUserPartition.

Examples

CreateStratifiedPartition(validationType = "CV", holdoutPct = 20, reps = 5)

Description

Creates a list object used by the SetTarget function to specify either Training/Validation/Holdout (validationType = "TVH") or cross-validation (validationType = "CV") partitions of the modeling dataset based on the values included in a column from the dataset. In either case, the name of this data column must be specified (as userPartitionCol).

Usage

CreateUserPartition(
  validationType,
  userPartitionCol,
  cvHoldoutLevel = NULL,
  trainingLevel = NULL,
  holdoutLevel = NULL,
  validationLevel = NULL
)

Arguments

Details

For the "TVH" option of cvMethod, no cross-validation is used. Users must specify the trainingLevel and validationLevel; use of a holdoutLevel is always recommended but not required. If no holdoutLevel is used, then the column must contain exactly 2 unique values. If a holdoutLevel is used, the column must contain exactly 3 unique values.

For the "CV" option, each value in the column will be used to separate rows into cross-validation folds. Use of a holdoutLevel is optional; if not specified, then no holdout is used.

This function is one of several convenience functions provided to simplify the task of starting modeling projects with custom partitioning options. The other functions are CreateGroupPartition, CreateRandomPartition, and CreateStratifiedPartition.

Value

An S3 object of class 'partition' including the parameters required by the SetTarget function to generate a user-specified of the modeling dataset.

See Also

CreateGroupPartition, CreateRandomPartition, CreateStratifiedPartition.

Examples

CreateUserPartition(validationType = "CV", userPartitionCol = "TVHflag", cvHoldoutLevel = NA)

Description

Note that this runs cross validation on a model as-is. If you would like to run cross-validation on a model with new parameters, use RequestNewModel instead.

Usage

CrossValidateModel(model)

Arguments

Details

Note that this is not implemented for prime models or datetime models.

Value

Job ID of the cross validation job.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  model <- GetModel(projectId, modelId)
  CrossValidateModel(model)

## End(Not run)

Description

This is a list that contains the valid values for CV methods

Usage

cvMethods

Format

An object of class list of length 5.

Description

This is a list that contains the valid values for data partitions

Usage

DataPartition

Format

An object of class list of length 3.

Description

Verifies that new data is either an existing datafile or a dataframe If a dataframe, save as a CSV file If neither an existing datafile nor a dataframe, halt with error

Usage

DataPathFromDataArg(dataSource, saveFile = NULL)

Arguments

Description

This is a list that contains the valid values for the dataSubset parameter found in RequestTrainingPredictions. If you wish, you can specify dataSubset using the list values here.

Usage

DataSubset

Format

An object of class list of length 4.

Details

For All, all available data is used.

For ValidationAndHoldout, only data outside the training set is used.

For Holdout, only holdout data is used.

For AllBacktests, data is used from all backtest validation folds. This requires the model to have successfully scored all backtests. Backtests are available on datetime partitioned projects only.

Description

Datetime trend plots resolutions

Usage

DatetimeTrendPlotsResolutions

Format

An object of class list of length 9.

Description

Datetime trend plots statuses

Usage

DatetimeTrendPlotsStatuses

Format

An object of class list of length 6.

Description

Record is deleted with preview and explanations.

Usage

DeleteAnomalyAssessmentRecord(projectId, recordId)

Arguments

See Also

Other Anomaly Assessment functions: GetAnomalyAssessmentExplanations(), GetAnomalyAssessmentPredictionsPreview(), InitializeAnomalyAssessment(), ListAnomalyAssessmentRecords()

Examples

## Not run: 
projectId <- "59a5af20c80891534e3c2bde"
recordId <- "59a5af20c80891534e3c2bdb"
explanations <- DeleteAnomalyAssessmentRecord(projectId, recordId)

## End(Not run)

Description

Delete a calendar

Usage

DeleteCalendar(calendarId)

Arguments

Examples

## Not run: 
   calendarId <- "5da75da31fb4a45b8a815a53"
   DeleteCalendar(calendarId)

## End(Not run)

Description

Note that default templates cannot be deleted.

Usage

DeleteComplianceDocTemplate(templateId)

Arguments

Value

Nothing returned, but deletes the compliance doc template.

Examples

## Not run: 
  templateId <- "5cf85080d9436e5c310c796d"
  DeleteComplianceDocTemplate(templateId)

## End(Not run)

Description

Delete a data store.

Usage

DeleteDataSource(dataSourceId)

Arguments

Examples

## Not run: 
 dataSourceId <- "5c1303269300d900016b41a7"
 DeleteDataSource(dataSourceId)

## End(Not run)

Description

Delete a data store.

Usage

DeleteDataStore(dataStoreId)

Arguments

Examples

## Not run: 
 dataStoreId <- "5c1303269300d900016b41a7"
 DeleteDataStore(dataStoreId)

## End(Not run)

Description

Delete a deployment.

Usage

DeleteDeployment(deploymentId)

Arguments

Examples

## Not run: 
  deploymentId <- "5e319d2e422fbd6b58a5edad"
  DeleteDeployment(deploymentId)

## End(Not run)

Description

Delete a featurelist

Usage

DeleteFeaturelist(featurelist)

Arguments

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  featureList <- CreateFeaturelist(projectId, "myFeaturelist", c("feature1", "feature2"))
  DeleteFeaturelist(featurelist)

## End(Not run)

Description

Cancel a running job

Usage

DeleteJob(job)

Arguments

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  initialJobs <- ListModelJobs(project)
  job <- initialJobs[[1]]
  DeleteJob(job)

## End(Not run)

Description

This function removes the model specified by the parameter model from its associated project.

Usage

DeleteModel(model)

Arguments

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  model <- GetModel(projectId, modelId)
  DeleteModel(model)

## End(Not run)

Description

Delete a modeling featurelist

Usage

DeleteModelingFeaturelist(featurelist)

Arguments

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  featureList <- CreateModelingFeaturelist(projectId, "myFeaturelist", c("feature1", "feature2"))
  featurelistId <- featureList$featurelistId
  GetModelingFeaturelist(projectId, featurelistId)
  DeleteModelingFeaturelist(projectId, featurelistId)

## End(Not run)

Description

This function deletes the modeling job specified by modelJobId from the DataRobot modeling queue.

Usage

DeleteModelJob(project, modelJobId)

Arguments

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  initialJobs <- ListModelJobs(project)
  job <- initialJobs[[1]]
  modelJobId <- job$modelJobId
  DeleteModelJob(projectId, modelJobId)

## End(Not run)

Description

This function removes a prediction dataset

Usage

DeletePredictionDataset(project, datasetId)

Arguments

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  datasets <- ListPredictionDatasets(projectId)
  dataset <- datasets[[1]]
  datasetId <- dataset$id
  DeletePredictionDataset(projectId, datasetId)

## End(Not run)

Description

This function deletes prediction explanations specified by project and predictionExplanationId.

Usage

DeletePredictionExplanations(project, predictionExplanationId)

Arguments

Value

Logical TRUE and displays a message to the user if the delete request was successful; otherwise an error message is displayed.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  datasets <- ListPredictionDatasets(projectId)
  dataset <- datasets[[1]]
  datasetId <- dataset$id
  model <- GetModel(projectId, modelId)
  jobId <- RequestPredictionExplanations(model, datasetId)
  predictionExplanationId <- GetPredictionExplanationsMetadataFromJobId(projectId, jobId)$id
  DeletePredictionExplanations(projectId, predictionExplanationId)

## End(Not run)

Description

Delete the prediction explanations initialization for a model.

Usage

DeletePredictionExplanationsInitialization(model)

Arguments

Value

Logical TRUE and displays a message to the user if the delete request was successful; otherwise an error message is displayed.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  model <- GetModel(projectId, modelId)
  DeletePredictionExplanationsInitialization(model)

## End(Not run)

Description

This function deletes the predict job specified by predictJobId from the DataRobot queue.

Usage

DeletePredictJob(project, predictJobId)

Arguments

Value

Logical TRUE and displays a message to the user if the delete request was successful; otherwise, execution halts and an error message is displayed.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  initialJobs <- GetPredictJobs(project)
  job <- initialJobs[[1]]
  predictJobId <- job$predictJobId
  DeletePredictJob(projectId, predictJobId)

## End(Not run)

Description

This function deletes the project defined by project, described under Arguments. This parameter may be obtained in several ways, including: (1), as one of the projectId elements of the list returned by ListProjects; (2), as the S3 object returned by the GetProject function; or (3), as the list returned by the SetupProject function.

Usage

DeleteProject(project)

Arguments

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  DeleteProject(projectId)

## End(Not run)

Description

Delete this imported model.

Usage

DeleteTransferableModel(importId)

Arguments

See Also

Other Transferable Model functions: DownloadTransferableModel(), GetTransferableModel(), ListTransferableModels(), RequestTransferableModel(), UpdateTransferableModel(), UploadTransferableModel()

Examples

## Not run: 
  id <- UploadTransferableModel("model.drmodel")
  DeleteTransferableModel(id)

## End(Not run)

Description

All possible deployment accuracy metrics. Added in DataRobot API 2.18.

Usage

DeploymentAccuracyMetric

Format

An object of class list of length 27.

Details

For usage, see DeploymentAccuracy and DeploymentAccuracyOverTime.

Description

Added in DataRobot API 2.18.

Usage

DeploymentServiceHealthMetric

Format

An object of class list of length 11.

Details

For usage, see GetDeploymentServiceStats.

Description

Differencing method

Usage

DifferencingMethod

Format

An object of class list of length 4.

Description

Download a compliance doc template (in JSON format).

Usage

DownloadComplianceDocTemplate(
  filename = "template.json",
  templateId = NULL,
  type = NULL
)

Arguments

Value

Nothing returned, but downloads the file to the stated filename.

Examples

## Not run: 
  DownloadComplianceDocTemplate("template.json")  # download the default template
  # download the default template
  DownloadComplianceDocTemplate("template.json", type = "normal")
  # download the default time series template
  DownloadComplianceDocTemplate("template.json" type = "timeSeries")
  templateId <- "5cf85080d9436e5c310c796d"
  DownloadComplianceDocTemplate(templateId) # Download a custom template for a specific ID.

## End(Not run)

Description

This function will create the compliance documentation first if it has not already been created. To create compliance documentation without downloading it, use CreateComplianceDocumentation. You can then skip the create step in this function by using 'create = FALSE'.

Usage

DownloadComplianceDocumentation(
  model,
  filename,
  templateId = NULL,
  create = TRUE,
  maxWait = 600
)

Arguments

Value

Nothing returned, but downloads the file to the stated filename.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  model <- GetModel(projectId, modelId)
  DownloadComplianceDocumentation(model)

## End(Not run)

Description

Function to download and save prediction explanations rows as csv file

Usage

DownloadPredictionExplanations(
  project,
  predictionExplanationId,
  filename,
  encoding = "UTF-8",
  excludeAdjustedPredictions = TRUE
)

Arguments

Value

Logical TRUE and displays a message to the user if the delete request was successful; otherwise an error message is displayed.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  datasets <- ListPredictionDatasets(projectId)
  dataset <- datasets[[1]]
  datasetId <- dataset$id
  model <- GetModel(projectId, modelId)
  jobId <- RequestPredictionExplanations(model, datasetId)
  predictionExplanationId <- GetPredictionExplanationsMetadataFromJobId(projectId, jobId)$id
  file <- file.path(tempdir(), "testPredictionExplanation.csv")
  DownloadPredictionExplanations(projectId, predictionExplanationId, file)

## End(Not run)

Description

Training a model using a ruleset is a necessary prerequisite for being able to download the code for a ruleset.

Usage

DownloadPrimeCode(project, primeFileId, filepath)

Arguments

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  primeFiles <- ListPrimeFiles(projectId)
  primeFile <- primeFiles[[1]]
  primeFileId <- primeFile$id
  file <- file.path(tempdir(), "primeCode.py")
  DownloadPrimeCode(projectId, primeFileId, file)

## End(Not run)

Description

Download a rating table to a CSV.

Usage

DownloadRatingTable(project, ratingTableId, filename)

Arguments

Value

Nothing returned, but downloads the file to the stated filename.

Examples

## Not run: 
 projectId <- "5984b4d7100d2b31c1166529"
 ratingTableId <- "5984b4d7100d2b31c1166529"
 file <- file.path(tempdir(), "ratingTable.csv")
 DownloadRatingTable(projectId, ratingTableId, file)

## End(Not run)

Description

Download scoring code JAR

Usage

DownloadScoringCode(project, modelId, fileName, sourceCode = FALSE)

Arguments

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  file <- file.path(tempdir(), "scoringCode.jar")
  DownloadScoringCode(projectId, modelId, file)

## End(Not run)

Description

Download the series accuracy for a model, computing it if not already computed.

Usage

DownloadSeriesAccuracy(model, filename, encoding = "UTF-8")

Arguments

Value

Nothing returned, but downloads the file to the stated filename.

Examples

## Not run: 
  projectId <- "5984b4d7100d2b31c1166529"
  modelId <- "5984b4d7100d2b31c1166529"
  model <- GetModel(projectId, modelId)
  DownloadSeriesAccuracy(model, "seriesAccuracy.csv")

## End(Not run)

Description

Download the time series feature derivation log as a text file.

Usage

DownloadTimeSeriesFeatureDerivationLog(project, file)

Arguments

Value

Nothing, but writes the output to the desired file.

See Also

GetTimeSeriesFeatureDerivationLog

Examples

## Not run: 
 projectId <- "5984b4d7100d2b31c1166529"
 DownloadTimeSeriesFeatureDerivationLog(projectId, "featureLog.txt")

## End(Not run)

Description

Download training predictions on a specified data set.

Usage

DownloadTrainingPredictions(
  project,
  predictionId,
  filename,
  encoding = "UTF-8"
)

Arguments

Value

NULL, but will produce a CSV with a dataframe with out-of-fold predictions for the training data.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  predictions <- ListTrainingPredictions(projectId)
  predictionId <- predictions[[1]]$predictionId
  file <- file.path(tempdir(), "myTrainingPredictions.csv")
  DownloadTrainingPredictions(projectId, predictionId, file)

## End(Not run)

Description

This function can only be used if model export is enabled, and will only be useful if you have an on-premise environment in which to import it.

Usage

DownloadTransferableModel(project, modelId, modelFile)

Arguments

See Also

Other Transferable Model functions: DeleteTransferableModel(), GetTransferableModel(), ListTransferableModels(), RequestTransferableModel(), UpdateTransferableModel(), UploadTransferableModel()

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  file <- file.path(tempdir(), "model.drmodel")
  DownloadTransferableModel(projectId, modelId, file)

## End(Not run)

Description

Make sure that the object has all of the keys specified. Also tests that there are not additional keys if allowAdditional is FALSE (default).

Usage

ExpectHasKeys(obj, keys, allowAdditional = FALSE)

Arguments

Description

If feature creation times out, the error message includes a URL corresponding to the creation task. That URL can be passed to this function (which will return the feature details when finished) to resume waiting for feature creation.

Usage

FeatureFromAsyncUrl(asyncUrl, maxWait = 600)

Arguments

Description

The DataRobot APIs expect dates formatted as RFC 3339 strings. This is the same as ISO 8601. To be safe, use UTC as the timezone (and format it with a 'Z' suffix), and use 'T' as the date/time separator.

Usage

formatRFC3339Timestamp(date)

Arguments

See Also

Other API datetime functions: RFC3339DateTimeFormat, parseRFC3339Timestamp(), transformRFC3339Period(), validateReportingPeriodTime()

Description

Based on the project dataset and the partitioning specification, inspect the full partitioning that would be used if the same specification were passed into SetTarget. This is not intended to be passed to SetTarget.

Usage

GenerateDatetimePartition(project, spec)

Arguments

Value

list describing datetime partition with following components

• cvMethod. The type of validation scheme used for the project.

• projectId character. The id of the project this partitioning applies to.

• datetimePartitionColumn character. The name of the column whose values as dates are used to assign a row to a particular partition.

• dateFormat character. The format (e.g. " partition column was interpreted (compatible with strftime [https://docs.python.org/2/library/time.html#time.strftime]).

• autopilotDataSelectionMethod character. Whether models created by the autopilot use "rowCount" or "duration" as their dataSelectionMethod.

• validationDuration character. The validation duration specified when initializing the partitioning - not directly significant if the backtests have been modified, but used as the default validationDuration for the backtests.

• availableTrainingStartDate character. The start date of the available training data for scoring the holdout.

• availableTrainingDuration character. The duration of the available training data for scoring the holdout.

• availableTrainingRowCount integer. The number of rows in the available training data for scoring the holdout. Only available when retrieving the partitioning after setting the target.

• availableTrainingEndDate character. The end date of the available training data for scoring the holdout.

• primaryTrainingStartDate character. The start date of primary training data for scoring the holdout.

• primaryTrainingDuration character. The duration of the primary training data for scoring the holdout.

• primaryTrainingRowCount integer. The number of rows in the primary training data for scoring the holdout. Only available when retrieving the partitioning after setting the target.

• primaryTrainingEndDate character. The end date of the primary training data for scoring the holdout.

• gapStartDate character. The start date of the gap between training and holdout scoring data.

• gapDuration character. The duration of the gap between training and holdout scoring data.

• gapRowCount integer. The number of rows in the gap between training and holdout scoring data. Only available when retrieving the partitioning after setting the target.

• gapEndDate character. The end date of the gap between training and holdout scoring data.

• holdoutStartDate character. The start date of holdout scoring data.

• holdoutDuration character. The duration of the holdout scoring data.

• holdoutRowCount integer. The number of rows in the holdout scoring data. Only available when retrieving the partitioning after setting the target.

• holdoutEndDate character. The end date of the holdout scoring data.

• numberOfBacktests integer. the number of backtests used.

• backtests data.frame. A data frame of partition backtest. Each element represent one backtest and has the following components: index, availableTrainingStartDate, availableTrainingDuration, availableTrainingRowCount, availableTrainingEndDate, primaryTrainingStartDate, primaryTrainingDuration, primaryTrainingRowCount, primaryTrainingEndDate, gapStartDate, gapDuration, gapRowCount, gapEndDate, validationStartDate, validationDuration, validationRowCount, validationEndDate, totalRowCount.

• useTimeSeries logical. Whether the project is a time series project (if TRUE) or an OTV project which uses datetime partitioning (if FALSE).

• defaultToKnownInAdvance logical. Whether the project defaults to treating features as known in advance. Known in advance features are time series features that are expected to be known for dates in the future when making predictions (e.g., "is this a holiday").

• featureDerivationWindowStart integer. Offset into the past to define how far back relative to the forecast point the feature derivation window should start. Only used for time series projects. Expressed in terms of the timeUnit of the datetimePartitionColumn.

• featureDerivationWindowEnd integer. Offset into the past to define how far back relative to the forecast point the feature derivation window should end. Only used for time series projects. Expressed in terms of the timeUnit of the datetimePartitionColumn.

• forecastWindowStart integer. Offset into the future to define how far forward relative to the forecast point the forecast window should start. Only used for time series projects. Expressed in terms of the timeUnit of the datetimePartitionColumn.

• forecastWindowEnd integer. Offset into the future to define how far forward relative to the forecast point the forecast window should end. Only used for time series projects. Expressed in terms of the timeUnit of the datetimePartitionColumn.

• featureSettings list. A list of lists specifying settings for each feature. For each feature you would like to set feature settings for, pass the following in a list:

  • featureName character. The name of the feature to set feature settings.

  • knownInAdvance logical. Optional. Whether or not the feature is known in advance. Used for time series only. Defaults to FALSE.

  • doNotDerive logical. Optional. If TRUE, no time series derived features (e.g., lags) will be automatically engineered from this feature. Used for time series only. Defaults to FALSE.

• treatAsExponential character. Specifies whether to treat data as exponential trend and apply transformations like log-transform. Uses values from from TreatAsExponential.

• differencingMethod character. Used to specify differencing method to apply if data is stationary. Use values from DifferencingMethod.

• windowsBasisUnit character. Indicates which unit is the basis for the feature derivation window and forecast window. Uses values from TimeUnit and the value "ROW".

• periodicities list. A list of periodicities for different times, specified as a list of lists, where each list item specifies the 'timeSteps' for a particular 'timeUnit'. Will be "ROW" if windowsBasisUnit is "ROW".

• totalRowCount integer. The number of rows in the project dataset. Only available when retrieving the partitioning after setting the target. Thus it will be NULL for GenerateDatetimePartition and populated for GetDatetimePartition.

• validationRowCount integer. The number of rows in the validation set.

• multiseriesIdColumns list. A list of the names of multiseries id columns to define series.

• numberOfKnownInAdvanceFeatures integer. The number of known in advance features.

• useCrossSeriesFeatures logical. Whether or not cross series features are included.

• aggregationType character. The aggregation type to apply when creating cross series features. See SeriesAggregationType.

• calendarId character. The ID of the calendar used for this project, if any.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  partitionSpec <- CreateDatetimePartitionSpecification("date_col")
  GenerateDatetimePartition(projectId, partitionSpec)

## End(Not run)

Description

Retrieve Accuracy over Time plot for a model.

Usage

GetAccuracyOverTimePlot(
  model,
  backtest = 0,
  source = SourceType$Validation,
  seriesId = NULL,
  forecastDistance = NULL,
  maxBinSize = NULL,
  resolution = NULL,
  startDate = NULL,
  endDate = NULL,
  maxWait = 600
)

Arguments

Value

list with the following components:

• resolution. character: The resolution that is used for binning. One of DatetimeTrendPlotsResolutions.

• startDate. POSIXct: The datetime of the start of the chartdata (inclusive).

• endDate. POSIXct: The datetime of the end of the chartdata (exclusive).

• bins. data.frame: Each row represents a bin in the plot. Dataframe has following columns:

  • startDate. POSIXct: The datetime of the start of the bin (inclusive).

  • endDate. POSIXct: The datetime of the end of the bin (exclusive).

  • actual. numeric: Average actual value of the target in the bin. NA if there are no entries in the bin.

  • predicted. numeric: Average prediction of the model in the bin. NA if there are no entries in the bin.

  • frequency. integer: Indicates number of values averaged in bin.

• statistics. list: Contains statistical properties for the plot.

  • durbinWatson. numeric: The Durbin-Watson statistic for the chart data. Value is between 0 and 4. Durbin-Watson statistic is a test statistic used to detect the presence of autocorrelation at lag 1 in the residuals (prediction errors) from a regression analysis.

• calendarEvents. data.frame: Each row represents a calendar event in the plot. Dataframe has following columns:

  • date. POSIXct: The date of the calendar event.

  • seriesId. character: The series ID for the event. If this event does not specify a series ID, then this will be NA, indicating that the event applies to all series.

  • name. character: The name of the calendar event.

Examples

## Not run: 
projectId <- "59a5af20c80891534e3c2bde"
modelId <- "5996f820af07fc605e81ead4"
model <- GetModel(projectId, modelId)
GetAccuracyOverTimePlot(model)
plot <- GetAccuracyOverTimePlot(model)
png("accuracy_over_time.png", width = 1200, height = 600, units = "px")
par(mar = c(10, 5, 5, 5))
plot(plot$bins$startDate, plot$bins$actual, type = "l", ylab = "Target", xaxt = "n", xlab = "")
lines(plot$bins$startDate, plot$bins$predicted, col = "red")
axis(1, plot$bins$startDate, format(plot$bins$startDate, "%Y-%m-%d"), las = 3)
title(xlab = "Date", mgp = c(7, 1, 0))
legend("topright", legend = c("Actual", "Predicted"), col = c("black", "red"), lty = 1:1)
dev.off()

## End(Not run)

Description

Retrieve Accuracy over Time preview plot for a model.

Usage

GetAccuracyOverTimePlotPreview(
  model,
  backtest = 0,
  source = SourceType$Validation,
  seriesId = NULL,
  forecastDistance = NULL,
  maxWait = 600
)

Arguments

Value

list with the following components:

• startDate. POSIXct: The datetime of the start of the chartdata (inclusive).

• endDate. POSIXct: The datetime of the end of the chartdata (exclusive).

• bins. data.frame: Each row represents a bin in the plot. Dataframe has following columns:

  • startDate. POSIXct: The datetime of the start of the bin (inclusive).

  • endDate. POSIXct: The datetime of the end of the bin (exclusive).

  • actual. numeric: Average actual value of the target in the bin. NA if there are no entries in the bin.

  • predicted. numeric: Average prediction of the model in the bin. NA if there are no entries in the bin.

Examples

## Not run: 
projectId <- "59a5af20c80891534e3c2bde"
modelId <- "5996f820af07fc605e81ead4"
model <- GetModel(projectId, modelId)
plot <- GetAccuracyOverTimePlotPreview(model)
png("accuracy_over_time_preview.png", width = 1200, height = 600, units = "px")
par(mar = c(10, 5, 5, 5))
plot(plot$bins$startDate, plot$bins$actual, type = "l", ylab = "Target", xaxt = "n", xlab = "")
lines(plot$bins$startDate, plot$bins$predicted, col = "red")
axis(1, plot$bins$startDate, format(plot$bins$startDate, "%Y-%m-%d"), las = 3)
title(xlab = "Date", mgp = c(7, 1, 0))
legend("topright", legend = c("Actual", "Predicted"), col = c("black", "red"), lty = 1:1)
dev.off()

## End(Not run)

Description

Retrieve Accuracy over Time plots metadata for a model.

Usage

GetAccuracyOverTimePlotsMetadata(model, forecastDistance = NULL)

Arguments

Value

list with the following components:

• forecastDistance. integer or NULL: The forecast distance for which the metadata was retrieved. NULL for OTV projects.

• resolutions. list: A list of DatetimeTrendPlotsResolutions, which represents available time resolutions for which plots can be retrieved.

• backtestStatuses. data.frame: Each row represents a status for the backtest SourceType. The row index corresponds to the backtest index via the relation rowIndex <- backtestIndex + 1. Status should be one of DatetimeTrendPlotsStatuses

• backtestMetadata. data.frame: Each row represents a metadata for the backtest SourceType start and end date. The row index corresponds to the backtest index via the relation rowIndex <- backtestIndex + 1. Each cell contains a POSIXct timestamp for start date (inclusive) and end date (exclusive) if the correspoding source type for the backtest is computed, and NA otherwise.

• holdoutStatuses. list: Contains statuses for holdout.

  • training. character: Status, one of DatetimeTrendPlotsStatuses

  • validation. character: Status, one of DatetimeTrendPlotsStatuses

• holdoutMetadata. list. Contains metadata for holdout.

  • training. list. Contains start and end date for holdout training.

  • validation. list. Contains start and end date for holdout validation.

    • startDate. POSIXct or NA: The datetime of the start of the holdout training/validation (inclusive). NA if the data is not computed.

    • endDate. POSIXct or NA: The datetime of the end of the holdout training/validation (exclusive). NA if the data is not computed.

Examples

## Not run: 
projectId <- "59a5af20c80891534e3c2bde"
modelId <- "5996f820af07fc605e81ead4"
model <- GetModel(projectId, modelId)
GetAccuracyOverTimePlotsMetadata(model)

## End(Not run)

Description

Explanations contain predictions along with shap explanations for the most anomalous records in the specified date range/for defined number of points. Two out of three parameters: startDate, endDate or pointsCount must be specified.

Usage

GetAnomalyAssessmentExplanations(
  projectId,
  recordId,
  startDate = NULL,
  endDate = NULL,
  pointsCount = NULL
)

Arguments

Value

The anomaly assessment explanations:

• recordId. character. The ID of the record.

• projectId. character. The project ID of the record.

• modelId. character. The model ID of the record.

• backtest. character. The backtest of the record.

• source. character. The source of the record.

• seriesId. character. the series ID of the record.

• startDate. POSIXct. First timestamp in the response. Will be NULL if there is no data in the specified range.

• endDate. POSIXct. Last timestamp in the response. Will be NULL if there is no data in the specified range.

• shapBaseValue. numeric. Shap base value.

• count. integer. The number of points in the data.

• data. list. A list of DataPoint objects in the specified date range containing:

  • shapExplanation. NULL or an array of up to 10 ShapleyFeatureContribution objects. Only rows with the highest anomaly scores have Shapley explanations calculated.

  • timestamp POSIXct. Timestamp for the row.

  • prediction numeric. The output of the model for this row.

  Each ShapleyFeatureContribution contains:
  

  • featureValue. character. The feature value for this row. First 50 characters are returned.

  • strength numeric. The shap value for this feature and row.

  • feature character. The feature name.

See Also

Other Anomaly Assessment functions: DeleteAnomalyAssessmentRecord(), GetAnomalyAssessmentPredictionsPreview(), InitializeAnomalyAssessment(), ListAnomalyAssessmentRecords()

Examples

## Not run: 
projectId <- "59a5af20c80891534e3c2bde"
recordId <- "59a5af20c80891534e3c2bdb"
explanations <- GetAnomalyAssessmentExplanations(projectId, recordId, pointsCount=100,
   startDate=as.Date("2021-01-01"))

## End(Not run)

Description

Aggregated predictions over time for the corresponding anomaly assessment record. Intended to find the bins with highest anomaly scores.

Usage

GetAnomalyAssessmentPredictionsPreview(projectId, recordId)

Arguments

Value

The anomaly assessment predictions preview:

• recordId. character. The ID of the record.

• projectId. character. The project ID of the record.

• modelId. character. The model ID of the record.

• backtest. character. The backtest of the record.

• source. character. The source of the record.

• seriesId. character. the series ID of the record.

• startDate. POSIXct. Timestamp of the first prediction in the subset.

• endDate. POSIXct. Timestamp of the last prediction in the subset.

• previewBins. list. A list of PreviewBin objects in the specified date range. The aggregated predictions for the subset. Bins boundaries may differ from actual start/end dates because this is an aggregation. Each PreviewBin contains:

  • startDate. POSIXct. Datetime of the start of the bin.

  • endDate. POSIXct. Datetime of the end of the bin.

  • avgPredicted numeric. The average prediction of the model in the bin. NA if there are no entries in the bin.

  • maxPredicted numeric. The maximum prediction of the model in the bin. NA if there are no entries in the bin.

  • frequency integer. The number of the rows in the bin.

See Also

Other Anomaly Assessment functions: DeleteAnomalyAssessmentRecord(), GetAnomalyAssessmentExplanations(), InitializeAnomalyAssessment(), ListAnomalyAssessmentRecords()

Examples

## Not run: 
projectId <- "59a5af20c80891534e3c2bde"
recordId <- "59a5af20c80891534e3c2bdb"
explanations <- GetAnomalyAssessmentPredictionsPreview(projectId, recordId)

## End(Not run)

Description

This function returns a DataRobot S3 object of class dataRobotModel for the model defined by project and modelId.

Usage

GetBlenderModel(project, modelId)

Arguments

Value

An S3 object of class ‘dataRobotBlenderModel’ summarizing all available information about the model. It is a list with the following components:

• modelId. character. The unique alphanumeric blender model identifier.

• modelNumber. integer. The assigned model number.

• modelType. character. The type of model, e.g. 'AVG Blender'.

• modelIds. character. List of unique identifiers for the blended models.

• blenderMethod. character. The blender method used to create this model.

• featurelistId. character. Unique alphanumeric identifier for the featurelist on which the model is based.

• processes. character. Components describing preprocessing; may include modelType.

• featurelistName. character. Name of the featurelist on which the model is based.

• blueprintId. character. The unique blueprint identifier on which the model is based.

• samplePct. numeric. The percentage of the dataset used in training the model. For projects that use datetime partitioning, this will be NA. See trainingRowCount instead.

• trainingRowCount. integer. Number of rows of the dataset used in training the model. For projects that use datetime partitioning, if specified, this defines the number of rows used to train the model and evaluate backtest scores; if unspecified, either trainingDuration or trainingStartDate and trainingEndDate was used instead.

• isFrozen. logical. Was the model created with frozen tuning parameters?

• metrics. list. The metrics associated with this model. Each element is a list with elements for each possible evaluation type (holdout, validation, and crossValidation).

• modelCategory. character. The category of model (e.g., blend, model, prime).

• projectId. character. Unique alphanumeric identifier for the project.

• projectName. character. Name of the project.

• projectTarget. character. The target variable predicted by all models in the project.

• projectMetric. character. The fitting metric optimized by all project models.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  GetBlenderModel(projectId, modelId)

## End(Not run)

Description

The function RequestBlender initiates the creation of new blender models in a DataRobot project.

Usage

GetBlenderModelFromJobId(project, modelJobId, maxWait = 600)

Arguments

Details

It submits requests to the DataRobot modeling engine and returns an integer-valued modelJobId. The GetBlenderModelFromJobId function polls the modeling engine until the model has been built or a specified time limit is exceeded, returning an S3 object of class 'dataRobotBlenderModel' when the model is available.

Motivation for this function is the fact that some models - e.g., very complex machine learning models fit to large datasets - may take a long time to complete. Splitting the model creation request from model retrieval in these cases allows the user to perform other interactive R session tasks between the time the model creation/update request is made and the time the final model is available.

Value

An S3 object of class ‘dataRobotBlenderModel’ summarizing all available information about the model. It is a list with the following components:

• modelId. character. The unique alphanumeric blender model identifier.

• modelNumber. integer. The assigned model number.

• modelType. character. The type of model, e.g. 'AVG Blender'.

• modelIds. character. List of unique identifiers for the blended models.

• blenderMethod. character. The blender method used to create this model.

• featurelistId. character. Unique alphanumeric identifier for the featurelist on which the model is based.

• processes. character. Components describing preprocessing; may include modelType.

• featurelistName. character. Name of the featurelist on which the model is based.

• blueprintId. character. The unique blueprint identifier on which the model is based.

• samplePct. numeric. The percentage of the dataset used in training the model. For projects that use datetime partitioning, this will be NA. See trainingRowCount instead.

• trainingRowCount. integer. Number of rows of the dataset used in training the model. For projects that use datetime partitioning, if specified, this defines the number of rows used to train the model and evaluate backtest scores; if unspecified, either trainingDuration or trainingStartDate and trainingEndDate was used instead.

• isFrozen. logical. Was the model created with frozen tuning parameters?

• metrics. list. The metrics associated with this model. Each element is a list with elements for each possible evaluation type (holdout, validation, and crossValidation).

• modelCategory. character. The category of model (e.g., blend, model, prime).

• projectId. character. Unique alphanumeric identifier for the project.

• projectName. character. Name of the project.

• projectTarget. character. The target variable predicted by all models in the project.

• projectMetric. character. The fitting metric optimized by all project models.

Examples

## Not run: 
projectId <- "59a5af20c80891534e3c2bde"
modelsToBlend <- c("5996f820af07fc605e81ead4", "59a5ce3301e9f0296721c64c")
blendJobId <- RequestBlender(projectId, modelsToBlend, "GLM")
GetBlenderModelFromJobId(projectId, blendJobId)

## End(Not run)

Description

Retrieve a blueprint

Usage

GetBlueprint(project, blueprintId)

Arguments

Value

List with the following four components:

projectId

Character string giving the unique DataRobot project identifier

processes

List of character strings, identifying any preprocessing steps included in the blueprint

blueprintId

Character string giving the unique DataRobot blueprint identifier

modelType

Character string, specifying the type of model the blueprint builds

blueprintCategory

Character string. Describes the category of the blueprint and the kind of model it produces.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  model <- GetModel(projectId, modelId)
  blueprintId <- model$blueprintId
  GetBlueprint(projectId, blueprintId)

## End(Not run)

Description

A Blueprint chart can be used to understand data flow in blueprint.

Usage

GetBlueprintChart(project, blueprintId)

Arguments

Value

List with the following two components:

• nodes. list each element contains information about one node of a blueprint : id and label.

• edges. Two column matrix, identifying blueprint nodes connections.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  model <- GetModel(projectId, modelId)
  blueprintId <- model$blueprintId
  GetBlueprintChart(projectId, blueprintId)

## End(Not run)

Description

Get documentation for tasks used in the blueprint

Usage

GetBlueprintDocumentation(project, blueprintId)

Arguments

Value

list with following components

task

Character string name of the task described in document

description

Character string task description

title

Character string title of document

parameters

List of parameters that task can received in human-readable format with following components: name, type, description

links

List of external lines used in document with following components: name, url

references

List of references used in document with following components: name, url

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  model <- GetModel(projectId, modelId)
  blueprintId <- model$blueprintId
  GetBlueprintDocumentation(projectId, blueprintId)

## End(Not run)

Description

Retrieve a calendar

Usage

GetCalendar(calendarId)

Arguments

Value

An S3 object of class "dataRobotCalendar"

Examples

## Not run: 
   calendarId <- "5da75da31fb4a45b8a815a53"
   GetCalendar(calendarId)

## End(Not run)

Description

Retrieve the calendar for a particular project.

Usage

GetCalendarFromProject(project)

Arguments

Value

An S3 object of class "dataRobotCalendar"

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
   GetCalendar(projectId)

## End(Not run)

Description

A custom compliance doc template can be retrieved using templateId. Default compliance doc templates that are built-in to DataRobot can be retrieved by using the type parameter. A type of NULL or "normal" will retrieve the default template. A type of "timeSeries" can be used to retrieve the default time series template.

Usage

GetComplianceDocTemplate(templateId = NULL, type = NULL)

Arguments

Value

An S3 object of class 'dataRobotComplianceDocTemplate' that contains:

• name character. The name of the compliance doc template.

• creatorUsername character. The name of the user who created the compliance doc template.

• orgId character. The ID of the organization of the creator user.

• creatorId character. The ID of the creator user.

• sections list. The list of sections that define the template.

• id character. The ID of the template.

Examples

## Not run: 
  GetComplianceDocTemplate()  # get the default template
  GetComplianceDocTemplate(type = "normal")  # get the default template
  GetComplianceDocTemplate(type = "timeSeries")  # get the default time series template
  templateId <- "5cf85080d9436e5c310c796d"
  GetComplianceDocTemplate(templateId) # Get a custom template for a specific ID.

## End(Not run)

Description

Retrieve a model's confusion chart for a specified source.

Usage

GetConfusionChart(
  model,
  source = DataPartition$VALIDATION,
  fallbackToParentInsights = FALSE
)

Arguments

Value

data.frame with the following components:

• source character. The name of the source of the confusion chart. Will be a member of DataPartition.

• data list. The data for the confusion chart, containing:

  • classes character. A vector containing the names of all the classes.

  • confusionMatrix matrix. A matrix showing the actual versus the predicted class values.

  • classMetrics list. A list detailing further metrics for each class:

    • wasActualPercentages data.frame. A dataframe detailing the actual percentage distribution of the classes.

    • wasPredictedPercentages data.frame. A dataframe detailing the predicted distribution of the classes.

    • f1 numeric. The F1 score for the predictions of the class.

    • recall numeric. The recall score for the predictions of the class.

    • precision numeric. The precision score for the predictions of the class.

    • actualCount integer. The actual count of values for the class.

    • predictedCount integer. The predicted count of values for the class.

    • className character. A vector containing the name of the class.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  GetModel(projectId, modelId)
  GetConfusionChart(modelId, source = DataPartition$VALIDATION)

## End(Not run)

Description

Get cross validation scores

Usage

GetCrossValidationScores(model, partition = NULL, metric = NULL)

Arguments

Value

A list of lists with cross validation score data. Each list contains a series of lists for each model metric. Each model metric list contains the metric data for each fold.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  model <- GetModel(projectId, modelId)
  GetCrossValidationScores(model)

## End(Not run)

Description

Returns information about a particular data source.

Usage

GetDataSource(dataSourceId)

Arguments

Value

A list containing information on the particular data source:

• className character. The Java class name of the driver.

• baseNames character. A vector of the file name(s) of the jar files.

• canonicalName character. The user-friendly name of the driver.

• id character. The dataSourceId of the driver.

• creator character. The userId of the user who created the driver.

Examples

## Not run: 
 dataSourceId <- "57a7c978c808916f4a630f89"
 GetDataSource(dataSourceId)

## End(Not run)

Description

Returns information about a particular data store.

Usage

GetDataStore(dataStoreId)

Arguments

Value

A list containing information on the particular data store:

• id character. The dataStoreId of the data store.

• canonicalName character. The user-friendly name of the data store.

• type character. The type of data store.

• updated datetime. A timestamp for the last time the data store was updated.

• creator character. The userId of the user who created the data store.

• params list. A list specifying the data store parameters.

Examples

## Not run: 
 dataStoreId <- "5c1303269300d900016b41a7"
 GetDataStore(dataStoreId)

## End(Not run)

Description

Get the schemas associated with a data store.

Usage

GetDataStoreSchemas(dataStoreId, username, password)

Arguments

Value

A list with the name of the catalog and the name of the schemas.

Examples

## Not run: 
 dataStoreId <- "5c1303269300d900016b41a7"
 GetDataStoreSchemas(dataStoreId, username = "myUser", password = "mySecurePass129")

## End(Not run)

Description

Get all tables associated with a data store.

Usage

GetDataStoreTables(dataStoreId, username, password, schema = NULL)

Arguments

Value

A list with the name of the catalog and the name of the tables.

Examples

## Not run: 
 dataStoreId <- "5c1303269300d900016b41a7"
 GetDataStoreTables(dataStoreId, username = "myUser", password = "mySecurePass129")

## End(Not run)

Description

This function returns a DataRobot S3 object of class dataRobotDatetimeModel for the model defined by project and modelId.

Usage

GetDatetimeModel(project, modelId)

Arguments

Details

If the project does not use datetime partitioning an error will occur.

Value

An S3 object of class 'dataRobotDatetimeModel', which is a list with the following components:

• featurelistId character. Unique alphanumeric identifier for the featurelist on which the model is based.

• processes character. Vector with components describing preprocessing; may include 'modelType'.

• featurelistName character. The name of the featurelist on which the model is based.

• projectId character. The unique alphanumeric identifier for the project.

• samplePct numeric. Percentage of the dataset used to form the training dataset for model fitting.

• isFrozen logical. Is model created with frozen tuning parameters?

• modelType character. A description of the model.

• metrics list. List with one element for each valid metric associated with the model. Each element is a list with elements for each possible evaluation type (holdout, validation, and crossValidation).

• modelCategory character. The model category (e.g., blend, model).

• blueprintId character. The unique DataRobot blueprint identifier on which the model is based.

• modelId character. The unique alphanumeric model identifier.

• modelNumber. integer. The assigned model number.

• projectName character. Optional description of project defined by projectId.

• projectTarget character. The target variable predicted by all models in the project.

• projectMetric character. The fitting metric optimized by all project models.

• trainingRowCount integer. The number of rows of the project dataset used in training the model. In a datetime partitioned project, if specified, defines the number of rows used to train the model and evaluate backtest scores; if unspecified, either trainingDuration or trainingStartDate and trainingEndDate was used to determine that instead.

• trainingDuration character. Only present for models in datetime partitioned projects. If specified, a duration string specifying the duration spanned by the data used to train the model and evaluate backtest scores.

• trainingStartDate character. Only present for frozen models in datetime partitioned projects. If specified, the start date of the data used to train the model.

• trainingEndDate character. Only present for frozen models in datetime partitioned projects. If specified, the end date of the data used to train the model.

• backtests list. What data was used to fit each backtest, the score for the project metric, and why the backtest score is unavailable if it is not provided.

• dataSelectionMethod character. Which of trainingRowCount, trainingDuration, or trainingStartDate and trainingEndDate were used to determine the data used to fit the model. One of "rowCount", "duration", or "selectedDateRange".

• trainingInfo list. Which data was used to train on when scoring the holdout and making predictions. trainingInfo will have the following keys: 'holdoutTrainingStartDate', 'holdoutTrainingDuration', 'holdoutTrainingRowCount', 'holdoutTrainingEndDate', 'predictionTrainingStartDate', 'predictionTrainingDuration', 'predictionTrainingRowCount', 'predictionTrainingEndDate'. Start and end dates will be datetime string, durations will be duration strings, and rows will be integers.

• holdoutScore numeric. The score against the holdout, if available and the holdout is unlocked, according to the project metric.

• holdoutStatus character. The status of the holdout score, e.g. "COMPLETED", "HOLDOUT_BOUNDARIES_EXCEEDED".

• effectiveFeatureDerivationWindowStart integer. Only available for time series projects. How many timeUnits into the past relative to the forecast point the user needs to provide history for at prediction time. This can differ from the 'featureDerivationWindowStart' set on the project due to the differencing method and period selected, or if the model is a time series native model such as ARIMA. Will be a negative integer in time series projects and 'NULL' otherwise.

• effectiveFeatureDerivationWindowEnd integer. Only available for time series projects. How many timeUnits into the past relative to the forecast point the feature derivation window should end. Will be a non-positive integer in time series projects and 'NULL' otherwise.

• forecastWindowStart integer. Only available for time series projects. How many timeUnits into the future relative to the forecast point the forecast window should start. Note that this field will be the same as what is shown in the project settings. Will be a non-negative integer in time series projects and 'NULL' otherwise.

• forecastWindowEnd integer. Only available for time series projects. How many timeUnits into the future relative to the forecast point the forecast window should end. Note that this field will be the same as what is shown in the project settings. Will be a non-negative integer in time series projects and 'NULL' otherwise.

• windowsBasisUnit character. Only available for time series projects. Indicates which unit is the basis for the feature derivation window and the forecast window. Note that this field will be the same as what is shown in the project settings. In time series projects, will be either the detected time unit or "ROW", and 'NULL' otherwise.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  modelId <- "5996f820af07fc605e81ead4"
  GetDatetimeModel(projectId, modelId)

## End(Not run)

Description

The functions RequestNewDatetimeModel and RequestFrozenDatetimeModel initiate the creation of new models in a DataRobot project. Both functions submit requests to the DataRobot modeling engine and return an integer-valued modelJobId. The GetDatetimeModelFromJobId function polls the modeling engine until the model has been built or a specified time limit is exceeded, returning an S3 object of class 'dataRobotDatetimeModel' when the model is available.

Usage

GetDatetimeModelFromJobId(project, modelJobId, maxWait = 600)

Arguments

Details

Motivation for this function is the fact that some models - e.g., very complex machine learning models fit to large datasets - may take a long time to complete. Splitting the model creation request from model retrieval in these cases allows the user to perform other interactive R session tasks between the time the model creation/update request is made and the time the final model is available.

Value

An S3 object of class 'dataRobotDatetimeModel' summarizing all available information about the model. See GetDatetimeModel

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  initialJobs <- ListModelJobs(project)
  job <- initialJobs[[1]]
  modelJobId <- job$modelJobId
  GetDatetimeModelFromJobId(projectId, modelJobId)

## End(Not run)

Description

Only available if the project has already set the target as a datetime project.

Usage

GetDatetimePartition(project)

Arguments

Value

list describing datetime partition with following components

• cvMethod. The type of validation scheme used for the project.

• projectId character. The id of the project this partitioning applies to.

• datetimePartitionColumn character. The name of the column whose values as dates are used to assign a row to a particular partition.

• dateFormat character. The format (e.g. " partition column was interpreted (compatible with strftime [https://docs.python.org/2/library/time.html#time.strftime]).

• autopilotDataSelectionMethod character. Whether models created by the autopilot use "rowCount" or "duration" as their dataSelectionMethod.

• validationDuration character. The validation duration specified when initializing the partitioning - not directly significant if the backtests have been modified, but used as the default validationDuration for the backtests.

• availableTrainingStartDate character. The start date of the available training data for scoring the holdout.

• availableTrainingDuration character. The duration of the available training data for scoring the holdout.

• availableTrainingRowCount integer. The number of rows in the available training data for scoring the holdout. Only available when retrieving the partitioning after setting the target.

• availableTrainingEndDate character. The end date of the available training data for scoring the holdout.

• primaryTrainingStartDate character. The start date of primary training data for scoring the holdout.

• primaryTrainingDuration character. The duration of the primary training data for scoring the holdout.

• primaryTrainingRowCount integer. The number of rows in the primary training data for scoring the holdout. Only available when retrieving the partitioning after setting the target.

• primaryTrainingEndDate character. The end date of the primary training data for scoring the holdout.

• gapStartDate character. The start date of the gap between training and holdout scoring data.

• gapDuration character. The duration of the gap between training and holdout scoring data.

• gapRowCount integer. The number of rows in the gap between training and holdout scoring data. Only available when retrieving the partitioning after setting the target.

• gapEndDate character. The end date of the gap between training and holdout scoring data.

• holdoutStartDate character. The start date of holdout scoring data.

• holdoutDuration character. The duration of the holdout scoring data.

• holdoutRowCount integer. The number of rows in the holdout scoring data. Only available when retrieving the partitioning after setting the target.

• holdoutEndDate character. The end date of the holdout scoring data.

• numberOfBacktests integer. the number of backtests used.

• backtests data.frame. A data frame of partition backtest. Each element represent one backtest and has the following components: index, availableTrainingStartDate, availableTrainingDuration, availableTrainingRowCount, availableTrainingEndDate, primaryTrainingStartDate, primaryTrainingDuration, primaryTrainingRowCount, primaryTrainingEndDate, gapStartDate, gapDuration, gapRowCount, gapEndDate, validationStartDate, validationDuration, validationRowCount, validationEndDate, totalRowCount.

• useTimeSeries logical. Whether the project is a time series project (if TRUE) or an OTV project which uses datetime partitioning (if FALSE).

• defaultToKnownInAdvance logical. Whether the project defaults to treating features as known in advance. Known in advance features are time series features that are expected to be known for dates in the future when making predictions (e.g., "is this a holiday").

• featureDerivationWindowStart integer. Offset into the past to define how far back relative to the forecast point the feature derivation window should start. Only used for time series projects. Expressed in terms of the timeUnit of the datetimePartitionColumn.

• featureDerivationWindowEnd integer. Offset into the past to define how far back relative to the forecast point the feature derivation window should end. Only used for time series projects. Expressed in terms of the timeUnit of the datetimePartitionColumn.

• forecastWindowStart integer. Offset into the future to define how far forward relative to the forecast point the forecast window should start. Only used for time series projects. Expressed in terms of the timeUnit of the datetimePartitionColumn.

• forecastWindowEnd integer. Offset into the future to define how far forward relative to the forecast point the forecast window should end. Only used for time series projects. Expressed in terms of the timeUnit of the datetimePartitionColumn.

• featureSettings list. A list of lists specifying settings for each feature. For each feature you would like to set feature settings for, pass the following in a list:

  • featureName character. The name of the feature to set feature settings.

  • knownInAdvance logical. Optional. Whether or not the feature is known in advance. Used for time series only. Defaults to FALSE.

  • doNotDerive logical. Optional. If TRUE, no time series derived features (e.g., lags) will be automatically engineered from this feature. Used for time series only. Defaults to FALSE.

• treatAsExponential character. Specifies whether to treat data as exponential trend and apply transformations like log-transform. Uses values from from TreatAsExponential.

• differencingMethod character. Used to specify differencing method to apply if data is stationary. Use values from DifferencingMethod.

• windowsBasisUnit character. Indicates which unit is the basis for the feature derivation window and forecast window. Uses values from TimeUnit and the value "ROW".

• periodicities list. A list of periodicities for different times, specified as a list of lists, where each list item specifies the 'timeSteps' for a particular 'timeUnit'. Will be "ROW" if windowsBasisUnit is "ROW".

• totalRowCount integer. The number of rows in the project dataset. Only available when retrieving the partitioning after setting the target. Thus it will be NULL for GenerateDatetimePartition and populated for GetDatetimePartition.

• validationRowCount integer. The number of rows in the validation set.

• multiseriesIdColumns list. A list of the names of multiseries id columns to define series.

• numberOfKnownInAdvanceFeatures integer. The number of known in advance features.

• useCrossSeriesFeatures logical. Whether or not cross series features are included.

• aggregationType character. The aggregation type to apply when creating cross series features. See SeriesAggregationType.

• calendarId character. The ID of the calendar used for this project, if any.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  GetDatetimePartition(projectId)

## End(Not run)

Description

Get information on a particular deployment.

Usage

GetDeployment(deploymentId)

Arguments

Value

A DataRobotDeployment object containing:

• id character. The ID of the deployment.

• label character. The label of the deployment.

• description character. The description of the deployment.

• defaultPredictionServer list. Information on the default prediction server connected with the deployment. See ListPredictionServers for details.

• model dataRobotModel. The model associated with the deployment. See GetModel for details.

• capabilities list. Information on the capabilities of the deployment.

• predictionUsage list. Information on the prediction usage of the deployment.

• permissions list. User's permissions on the deployment.

• serviceHealth list. Information on the service health of the deployment.

• modelHealth list. Information on the model health of the deployment.

• accuracyHealth list. Information on the accuracy health of the deployment.

Examples

## Not run: 
  deploymentId <- "5e319d2e422fbd6b58a5edad"
  GetDeployment(deploymentId)

## End(Not run)

Description

Retrieve accuracy statistics for a deployment.

Usage

GetDeploymentAccuracy(
  deploymentId,
  modelId = NULL,
  start = NULL,
  end = NULL,
  segmentAttribute = NULL,
  segmentValue = NULL,
  targetClasses = NULL
)

Arguments

Value

An object representing service health metrics for the deployment, containing:

• modelId character. The ID of the deployment model for which monitoring data was retrieved.

• period list. The duration of the reporting period, containing:

  • start POSIXct. Start of the reporting period.

  • end POSIXct. End of the reporting period.

• metrics data.frame. Accuracy metrics for the deployment, where each row is a separate metric and contains the columns:

  • metric. character. Name of the metric. See DeploymentAccuracyMetric for valid values.

  • baselineValue. numeric. May be NA if accuracy data is not available.

  • value. numeric. May be NA if accuracy data is not available.

  • percentChange. numeric. The percent change of value over baseline. May be NA if accuracy data is not available.

• segmentAttribute character. Optional. The name of the segment on which segment analysis was performed. Added in DataRobot 2.21.

• segmentValue character. Optional. The value of the segmentAttribute. Added in DataRobot 2.21.

See Also

Other deployment accuracy functions: GetDeploymentAccuracyOverTime(), GetDeploymentAssociationId(), SubmitActuals()

Examples

## Not run: 
library(dplyr)
deploymentId <- "59a5af20c80891534e3c2bde"
acc <- GetDeploymentAccuracy(deploymentId, end = ISOdate(2021, 01, 06, 1, 0, 0, tz = "UTC"))
df <- mutate(
       acc$metrics,
       "modelId" = acc$modelId,
       "startTime" = acc$period$start,
       "endTime" = acc$period$end,
       .before = everything()
   )

## End(Not run)

Description

By default this will return statistics for the last seven days prior to the next; set the start and end parameters to adjust the reporting period.

Usage

GetDeploymentAccuracyOverTime(
  deploymentId,
  metrics,
  modelId = NULL,
  start = NULL,
  end = NULL,
  bucketSize = NULL,
  segmentAttribute = NULL,
  segmentValue = NULL
)

Arguments

Value

An object representing how accuracy has changed over time for the deployment, containing:

• modelId character. The ID of the deployment model for which monitoring data was retrieved.

• summary data.frame. A summary bucket across the entire reporting period.

• buckets data.frame. A list of buckets representing each interval (constrained by the bucketSize parameter) in the reporting period.

• baseline data.frame. A baseline bucket.

Each bucket contains:

• sampleSize. integer. The number of predictions made against this deployment.

• start. POSIXct. The start time of the bucket. May be NA.

• end. POSIXct. The end time of the bucket. May be NA.

• metricName. numeric. Given N metrics queried, there will be N value columns, each one named for the metric. See DeploymentAccuracyMetric for supported values. May be NA if sampleSize is 0.

See Also

Other deployment accuracy functions: GetDeploymentAccuracy(), GetDeploymentAssociationId(), SubmitActuals()

Examples

## Not run: 
deploymentId <- "59a5af20c80891534e3c2bde"
aot <- GetDeploymentAccuracyOverTime(deploymentId,
         metrics = c(DeploymentAccuracyMetric$Gamma.Deviance,
                     DeploymentAccuracyMetric$LogLoss,
                     DeploymentAccuracyMetric$RMSE))

## End(Not run)

Description

The association ID of a deployment is a foreign key for your prediction dataset that will be used to match up actual values with those predictions. The ID should correspond to an event for which you want to track the outcome.

Usage

GetDeploymentAssociationId(deployment)

UpdateDeploymentAssociationId(
  deployment,
  columnNames = c(),
  requiredInPredictionRequests = NULL,
  maxWait = 600
)

Arguments

Details

These functions are convenience methods to get and set the association ID settings for a deployment.

Value

An object classed dataRobotDeploymentAssociationIdSettings that contains:

columnNames

character. The columns that can be used as association IDs.

requiredInPredictionRequests

logical. Whether the association ID is required in a prediction request.

Functions

• UpdateDeploymentAssociationId(): Updates the association ID settings of a deployment. It will only update those settings that correspond to set arguments. This function will throw an error if the update fails and return the updated settings on success.

See Also

Other deployment accuracy functions: GetDeploymentAccuracyOverTime(), GetDeploymentAccuracy(), SubmitActuals()

Description

Get drift tracking settings for a deployment.

Usage

GetDeploymentDriftTrackingSettings(deploymentId)

Arguments

Value

A list with the following information on drift tracking:

• associationId

• predictionIntervals list. A list with two keys:

  • enabled. 'TRUE' if prediction intervals are enabled and 'FALSE' otherwise.

  • percentiles list. A list of percentiles, if prediction intervals are enabled.

• targetDrift list. A list with one key, 'enabled', which is 'TRUE' if target drift is enabled, and 'FALSE' otherwise.

• featureDrift list. A list with one key, 'enabled', which is 'TRUE' if feature drift is enabled, and 'FALSE' otherwise.

Examples

## Not run: 
  deploymentId <- "5e319d2e422fbd6b58a5edad"
  GetDeploymentDriftTrackingSettings(deploymentId)

## End(Not run)

Description

Retrieve service health statistics for a deployment.

Usage

GetDeploymentServiceStats(
  deploymentId,
  modelId = NULL,
  start = NULL,
  end = NULL,
  executionTimeQuantile = NULL,
  responseTimeQuantile = NULL,
  slowRequestsThreshold = NULL,
  segmentAttribute = NULL,
  segmentValue = NULL
)

Arguments

Value

An object representing service health metrics for the deployment, containing:

• modelId character. The ID of the deployment model for which monitoring data was retrieved.

• period list. The duration of the reporting period, containing:

  • start POSIXct. Start of the reporting period.

  • end POSIXct. End of the reporting period.

• metrics list. Service health metrics for the deployment, containing:

  • totalPredictions integer. Total number of prediction rows.

  • totalRequests integer. Total number of prediction requests performed.

  • slowRequests integer. Number of requests with response time greater than slowRequestsThreshold.

  • responseTime numeric. Request response time at responseTimeQuantile in milliseconds. May be NA.

  • executionTime numeric. Request execution time at executionTimeQuantile in milliseconds. May be NA.

  • medianLoad integer. Median request rate, in requests per minute.

  • peakLoad integer. Greatest request rate, in requests per minute.

  • userErrorRate numeric. Ratio of user errors to the total number of requests.

  • serverErrorRate numeric. Ratio of server errors to the total number of requests.

  • numConsumers integer. Number of unique users performing requests.

  • cacheHitRatio numeric. The ratio of cache hits to requests.

• segmentAttribute character. Added in DataRobot 2.20. The name of the segment on which segment analysis was performed.

• segmentValue character. Added in DataRobot 2.20. The value of the segmentAttribute.

Examples

## Not run: 
  deploymentId <- "59a5af20c80891534e3c2bde"
  startTime = ISOdate(2020, 12, 25, 1, 0, 0, tz = "UTC")
  endTime = ISOdate(2021, 01, 06, 1, 0, 0, tz = "UTC")
  GetDeploymentServiceStats(deploymentId, startTime, endTime)

## End(Not run)
## Not run: 
  deploymentId <- "59a5af20c80891534e3c2bde"
  GetDeploymentServiceStats(deploymentId,
                            segmentAttribute = SegmentAnalysisAttribute$DataRobotRemoteIP,
                            segmentValue = "192.168.0.1")

## End(Not run)

Description

By default this will return statistics for the last seven days prior to the next; set the start and end parameters to adjust the reporting period.

Usage

GetDeploymentServiceStatsOverTime(
  deploymentId,
  metrics = DeploymentServiceHealthMetric$TotalPredictions,
  modelId = NULL,
  start = NULL,
  end = NULL,
  bucketSize = NULL,
  quantile = NULL,
  threshold = NULL,
  segmentAttribute = NULL,
  segmentValue = NULL
)

Arguments

Value

• modelId character. The ID of the deployment model for which monitoring data was retrieved.

• summary data.frame. Summarizes statistics for each metric over the entire reporting period.

• buckets data.frame. Statistics for each metric, split into intervals of equal duration. There is one column representing stats for each metric queried, as well as:

  • start POSIXct. Start of the interval.

  • end POSIXct. End of the interval.

• segmentAttribute character. Added in DataRobot 2.20. The name of the segment on which segment analysis was performed.

• segmentValue character. Added in DataRobot 2.20. The value of segmentAttribute.

Examples

## Not run: 
metrics <- c(DeploymentServiceHealthMetric)
GetDeploymentServiceStatsOverTime(deploymentId, metrics = metrics)

## End(Not run)

Description

Returns information about a particular driver.

Usage

GetDriver(driverId)

Arguments

Value

A list containing information on the particular driver:

• className character. The Java class name of the driver.

• baseNames character. A vector of the file name(s) of the jar files.

• canonicalName character. The user-friendly name of the driver.

• id character. The driverId of the driver.

• creator character. The userId of the user who created the driver.

Examples

## Not run: 
 driverId <- "57a7c978c808916f4a630f89"
 GetDriver(driverId)

## End(Not run)

Description

Get pairwise feature association statistics for a project's informative features

Usage

GetFeatureAssociationMatrix(project, associationType, metric)

Arguments

Value

A list with two items:

• features data.frame. A data.frame containing the following info for each feature:

  • alphabeticSortIndex integer. A number representing the alphabetical order of this feature compared to the other features in this dataset.

  • feature character. The name of the feature.

  • importanceSortIndex integer. A number ranking the importance of this feature compared to the other features in this dataset.

  • strengthSortIndex integer. A number ranking the strength of this feature compared to the other features in this dataset.

• strengths data.frame. A data.frame of pairwise strength data, with the following info:

  • feature1 character. The name of the first feature.

  • feature2 character. The name of the second feature.

  • statistic numeric. Feature association statistics for 'feature1' and 'feature2'.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  GetFeatureAssociationMatrix(projectId)

## End(Not run)

Description

Get a sample of the actual values used to measure the association between a pair of features.

Usage

GetFeatureAssociationMatrixDetails(project, feature1, feature2)

Arguments

Value

A list with the following info:

• features list. The names of 'feature1' and 'feature2'.

• types list. The type of 'feature1' and 'feature2'. Will be "C" for categorical and "N" for numeric.

• values data.frame. The values of the feature associations and the relative frequency of the data points in the sample.

Examples

## Not run: 
  projectId <- "59a5af20c80891534e3c2bde"
  GetFeatureAssociationMatrix(projectId, "SepalWidth", "SepalLength")

## End(Not run)

Description

A histogram is a popular way of visual representation of a feature values distribution in a series of bins. For categorical features every bin represents exactly one of feature values plus the number of occurrences of that value. For numeric features every bin represents a range of values (low end inclusive, high end exclusive) plus the total number of occurrences of all values in this range. In addition to that, with every bin for categorical and numeric features there is also included a target feature average for values in that bin (though it can be missing if the feature is deemed uninformative, if the project target has not been selected yet using SetTarget, or if the project is a multiclass project).

Usage

GetFeatureHistogram(project, featureName, binLimit = NULL)

Arguments

Value

list containing:

• count numeric. The number of values in this bin's range. If a project is using weights, the value is equal to the sum of weights of all feature values in the bin's range.

• target numeric. Average of the target feature for values in this bin. It may be NULL if the feature is deemed uninformative, if the target has not yet been set (see SetTarget), or if the project is multiclass.

• label character. The value of the feature if categorical, otherwise the low end of the bin range such that the difference between two consecutive bin labels is the length of the bin.

Description

Feature Impact is computed for each column by creating new data with that column randomly permuted (but the others left unchanged), and seeing how the error metric score for the predictions is affected. The 'impactUnnormalized' is how much worse the error metric score is when making predictions on this modified data. The 'impactNormalized' is normalized so that the largest value is 1. In both cases, larger values indicate more important features. Elsewhere this technique is sometimes called 'Permutation Importance'.

Usage

GetFeatureImpact(model)

Arguments

Details

Note that GetFeatureImpact will block for the duration of feature impact calculation. If you would prefer not to block the call, use RequestFeatureImpact to generate an async request for feature impact and then use GetFeatureImpactForModel or GetFeatureImpactForJobId to get the feature impact when it has been calculated. GetFeatureImpactForJobId will also block until the request is complete, whereas GetFeatureImpactForModel will error if the job is not complete yet.

Description

This will wait for the Feature Impact job to be completed (giving an error if the job is not a Feature Impact job and an error if the job errors).

Usage

GetFeatureImpactForJobId(project, jobId, maxWait = 600)

Arguments

Value

A data frame with the following columns:

• featureName character. The name of the feature.

• impactNormalized numeric. The normalized impact score (largest value is 1).

• impactUnnormalized numeric. The unnormalized impact score.

• redundantWith character. A feature that makes this feature redundant, or NA if the feature is not redundant.

Examples

## Not run: 
  model <- ListModels(project)[[1]]
  featureImpactJobId <- RequestFeatureImpact(model)
  featureImpact <- GetFeatureImpactForJobId(project, featureImpactJobId)

## End(Not run)

Description

This will only succeed if the Feature Impact computation has completed.

Usage

GetFeatureImpactForModel(model)

Arguments

Details

Feature Impact is computed for each column by creating new data with that column randomly permuted (but the others left unchanged), and seeing how the error metric score for the predictions is affected. The 'impactUnnormalized' is how much worse the error metric score is when making predictions on this modified data. The 'impactNormalized' is normalized so that the largest value is 1. In both cases, larger values indicate more important features. Elsewhere this technique is sometimes called 'Permutation Importance'.

Feature impact also runs redundancy detection, which detects if some features are redundant with higher importance features. Note that some types of projects, like multiclass, do not run redundancy detection. This function will generate a warning if redundancy detection was not run.

Value

A data frame with the following columns:

• featureName character. The name of the feature.

• impactNormalized numeric. The normalized impact score (largest value is 1).

• impactUnnormalized numeric. The unnormalized impact score.

• redundantWith character. A feature that makes this feature redundant, or NA if the feature is not redundant.

Examples

## Not run: 
  model <- ListModels(project)[[1]]
  featureImpactJobId <- RequestFeatureImpact(model)
  # Note: This will only work after the feature impact job has completed. Use
  #       GetFeatureImpactFromJobId to automatically wait for the job.\
  featureImpact <- GetFeatureImpactForModel(model)

## End(Not run)

Description

Details about a feature

Usage

GetFeatureInfo(project, featureName)

Arguments

Value

A named list which contains:

• id numeric. feature id. Note that throughout the API, features are specified using their names, not this ID.

• name character. The name of the feature.

• featureType character. Feature type: 'Numeric', 'Categorical', etc.

• importance numeric. numeric measure of the strength of relationship between the feature and target (independent of any model or other features).

• lowInformation logical. Whether the feature has too few values to be informative.

• uniqueCount numeric. The number of unique values in the feature.

• naCount numeric. The number of missing values in the feature.

• dateFormat character. The format of the feature if it is date-time feature.

• projectId character. Character id of the project the feature belongs to.

• max. The maximum value in the dataset, formatted in the same format as the data.

• min. The minimum value in the dataset, formatted in the same format as the data.

• mean. The arithmetic mean of the dataset, formatted in the same format as the data.

• median. The median of the dataset, formatted in the same format as the data.

• stdDev. The standard deviation of the dataset, formatted in the same format as the data.

• timeSeriesEligible logical. Whether this feature can be used as the datetime partition column in a time series project.

• timeSeriesEligibilityReason character. Why the feature is ineligible for the datetime partition column in a time series project, "suitable" when it is eligible.

• crossSeriesEligible logical. Whether the cross series group by column is eligible for cross-series modeling. Will be NULL if no cross series group by column is used.

• crossSeriesEligibilityReason character. The type of cross series eligibility (or ineligibility).

• timeStep numeric. For time-series eligible features, a positive integer determining the interval at which windows can be specified. If used as the datetime partition column on a time series project, the feature derivation and forecast windows must start and end at an integer multiple of this value. NULL for features that are not time series eligible.

• timeUnit character. For time series eligible features, the time unit covered by a single time step, e.g. "HOUR", or NULL for features that are not time series eligible.

• targetLeakage character. Whether a feature is considered to have target leakage or not. A value of "SKIPPED_DETECTION" indicates that target leakage detection was not run on the feature.

• keySummary data.frame. Optional. Descriptive statistics for this feature, iff it is a summarized categorical feature. This data.frame contains:

  • key. The name of the key.

  • summary. Descriptive statistics for this key, including:

    • max. The maximum value in the dataset.

    • min. The minimum value in the dataset.

    • mean. The arithmetic mean of the dataset.

    • median. The median of the dataset.

    • stdDev. The standard deviation of the dataset.

    • pctRows. The percentage of rows (from the EDA sample) in which this key occurs.