| Title: | Graphical User Interface for Data Science in R |
|---|---|
| Description: | The R Analytic Tool To Learn Easily (Rattle) provides a collection of utilities functions for the data scientist. A Gnome (RGtk2) based graphical interface is included with the aim to provide a simple and intuitive introduction to R for data science, allowing a user to quickly load data from a CSV file (or via ODBC), transform and explore the data, build and evaluate models, and export models as PMML (predictive modelling markup language) or as scores. A key aspect of the GUI is that all R commands are logged and commented through the log tab. This can be saved as a standalone R script file and as an aid for the user to learn R or to copy-and-paste directly into R itself. Note that RGtk2 and cairoDevice have been archived on CRAN. See <https://rattle.togaware.com> for installation instructions. |
| Authors: | Graham Williams [aut, cph, cre], Mark Vere Culp [cph], Ed Cox [ctb], Anthony Nolan [ctb], Denis White [cph], Daniele Medri [ctb], Akbar Waljee [ctb] (OOB AUC for Random Forest), Brian Ripley [cph] (print.summary.nnet), Jose Magana [ctb] (ggpairs plots), Surendra Tipparaju [ctb] (initial RevoScaleR/XDF), Durga Prasad Chappidi [ctb] (initial RevoScaleR/XDF), Dinesh Manyam Venkata [ctb] (initial RevoScaleR/XDF), Mrinal Chakraborty [ctb] (initial RevoScaleR/XDF), Fang Zhou [ctb] (initial xgboost), Cameron Chisholm [ctb] (risk plot on risk chart) |
| Maintainer: | Graham Williams <[email protected]> |
| License: | GPL (>= 2) |
| Version: | 5.5.1 |
| Built: | 2024-11-08 06:50:06 UTC |
| Source: | CRAN |
Rattle uses an artificial dataset for demonstration purposes. This function retrieves the source data https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.data and then transforms the data in a variety of ways.
acquireAuditData(write.to.file=FALSE)acquireAuditData(write.to.file=FALSE)
write.to.file |
Whether to generate a colleciton of files based on the data. The files generated include: audit.csv, audit.Rdata, audit.arf, and audit\_missing.csv |
See the function definition for details of the processing done on the data downloaded from the UCI repository.
By default the function returns a data frame containing the audit dataset. If write.to.file is TRUE then the data frame is returned invisibly.
Package home page: https://rattle.togaware.com
Display a list of rules for an rpart decision tree.
asRules(model, compact=FALSE, ...)asRules(model, compact=FALSE, ...)
model |
an rpart model. |
compact |
whether to list cateogricals compactly. |
... |
further arguments passed to or from other methods. |
Traverse a decision tree to generate the equivalent set of rules, one rule for each path from the root node to a leaf node.
Package home page: https://rattle.togaware.com
## Not run: asRules.rpart(my.rpart)## Not run: asRules.rpart(my.rpart)
Display a list of rules for an rpart decision tree.
## S3 method for class 'rpart' asRules(model, compact=FALSE, classes=NULL, ...)## S3 method for class 'rpart' asRules(model, compact=FALSE, classes=NULL, ...)
model |
an rpart model. |
compact |
whether to list cateogricals compactly (default FALSE). |
classes |
which target classes should be listed (default all). |
... |
further arguments passed to or from other methods. |
Traverse a decision tree to generate the equivalent set of rules, one rule for each path from the root node to a leaf node.
Package home page: https://rattle.togaware.com
## Not run: asRules.rpart(my.rpart)## Not run: asRules.rpart(my.rpart)
The audit dataset is an artificially constructed dataset that has some of the characteristics of a true financial audit dataset for modelling productive and non-productive audits of a person's financial statement. A productive audit is one which identifies errors or inaccuracies in the information provided by a client. A non-productive audit is usually an audit which found all supplied information to be in order.
The audit dataset is used to illustrate binary classification. The
target variable is identified as TARGET\_Adjusted.
The dataset is quite small, consisting of just 2000 entities. Its primary purpose is to illustrate modelling in Rattle, so a minimally sized dataset is suitable.
The dataset itself is derived from publicly available data (which has nothing to do with audits).
A data frame. In line with data mining terminology we refer to the rows of the data frame (or the observations) as entities. The columns are refered to as variables. The entities represent people in this case. We describe the variables here:
IDThis is a unique identifier for each person.
AgeThe age.
EmploymentThe type of employment.
EducationThe highest level of education.
MaritalCurrent marital status.
OccupationThe type of occupation.
IncomeThe amount of income declared.
GenderThe persons gender.
DeductionsTotal amount of expenses that a person claims in their financial statement.
HoursThe average hours worked on a weekly basis.
IGNORE_AccountsThe main country in which the person has most of their money banked. Note that the variable name is prefixed with IGNORE. This is recognised by Rattle as the default role for this variable.
RISK_AdjustmentThis variable records the monetary amount of any adjustment to the person's financial claims as a result of a productive audit. This variable, which should not be treated as an input variable, is thus a measure of the size of the risk associated with the person.
TARGET_AdjustedThe target variable for modelling (generally for classification modelling). This is a numeric field of class integer, but limited to 0 and 1, indicating non-productive and productive audits, respectively. Productive audits are those that result in an adjustment being made to a client's financial statement.
Perform binning.
binning(x, bins=4, method=c("quantile", "wtd.quantile", "kmeans"), labels=NULL, ordered=TRUE, weights=NULL)binning(x, bins=4, method=c("quantile", "wtd.quantile", "kmeans"), labels=NULL, ordered=TRUE, weights=NULL)
x |
the numeric data to bin. |
bins |
the number of bins to use. |
method |
whether to use "quantile", weighted quantile "wtd.quantile" or "kmeans" binning. |
labels |
the labels or names to use for each of the bins. |
ordered |
whether to build an ordered factor or not. |
weights |
vector of numeric weights for each observation for weighted quantile binning. |
Bin the provided nmeric data into the specified number of bins using one of the supported methods. The bins will have the names specified by labels, if supplied. The result can optionally be an ordered factor.
A factor is returned.
Daniele Medri and Graham Williams
Package home page: https://rattle.togaware.com
In the context of Benford's Law calculate the distribution of the frequencies of the first digit of the numbers supplied as the argument.
calcInitialDigitDistr(l, digit=1, len=1, sp=c("none", "positive", "negative"))calcInitialDigitDistr(l, digit=1, len=1, sp=c("none", "positive", "negative"))
l |
a vector of numbers. |
digit |
the digit to generate frequencies for. |
len |
The number of digits. |
sp |
whether and how to split the digits. |
Package home page: https://rattle.togaware.com
Given the evaluation returned by evaluateRisk, for example, calculate the area under the risk or recall curves, to use as a metric to compare the performance of a model.
calculateAUC(x, y)calculateAUC(x, y)
x |
a vector of values for the x points. |
y |
a vector of values for the y points. |
The area is returned.
Package home page: https://rattle.togaware.com
## this is usually used in the context of the evaluateRisk function ## Not run: ev <- evaluateRisk(predicted, actual, risk) ## imitate this output here ev <- data.frame(Caseload=c(1.0, 0.8, 0.6, 0.4, 0.2, 0), Precision=c(0.15, 0.18, 0.21, 0.25, 0.28, 0.30), Recall=c(1.0, 0.95, 0.80, 0.75, 0.5, 0.0), Risk=c(1.0, 0.98, 0.90, 0.77, 0.30, 0.0)) ## Calculate the areas unde the Risk and the Recall curves. calculateAUC(ev$Caseload, ev$Risk) calculateAUC(ev$Caseload, ev$Recall)## this is usually used in the context of the evaluateRisk function ## Not run: ev <- evaluateRisk(predicted, actual, risk) ## imitate this output here ev <- data.frame(Caseload=c(1.0, 0.8, 0.6, 0.4, 0.2, 0), Precision=c(0.15, 0.18, 0.21, 0.25, 0.28, 0.30), Recall=c(1.0, 0.95, 0.80, 0.75, 0.5, 0.0), Risk=c(1.0, 0.98, 0.90, 0.77, 0.30, 0.0)) ## Calculate the areas unde the Risk and the Recall curves. calculateAUC(ev$Caseload, ev$Risk) calculateAUC(ev$Caseload, ev$Recall)
Generate a matrix of centers from a hierarchical cluster.
centers.hclust(x, object, nclust=10, use.median=FALSE)centers.hclust(x, object, nclust=10, use.median=FALSE)
x |
The data used to build the cluster. |
object |
A hclust object. |
nclust |
Number of clusters. |
use.median |
Use meadion instead of mean. |
For the specified number of clusters, cut the hierarchical cluster appropriately to that number of clusters, and return the mean (or median) of each resulting cluster.
Daniele Medri and [email protected]
Package home page: https://rattle.togaware.com
Format data in the most appropriate human readable form.
comcat(x, ...)comcat(x, ...)
x |
object. |
... |
additional arguments passed on to format. |
Package home page: https://rattle.togaware.com
comcat(dim(iris))comcat(dim(iris))
Draw the nodes of a decision tree
drawTreeNodes(tree, cex = par("cex"), pch = par("pch"), size = 4 * cex, col = NULL, nodeinfo = FALSE, units = "", cases = "obs", digits = getOption("digits"), decimals = 2, print.levels = TRUE, new = TRUE)drawTreeNodes(tree, cex = par("cex"), pch = par("pch"), size = 4 * cex, col = NULL, nodeinfo = FALSE, units = "", cases = "obs", digits = getOption("digits"), decimals = 2, print.levels = TRUE, new = TRUE)
tree |
an rpart decision tree. |
cex |
. |
pch |
. |
size |
. |
col |
. |
nodeinfo |
. |
units |
. |
cases |
. |
digits |
. |
decimals |
the number of decimal digits to include in numeric split nodes. |
print.levels |
. |
new |
. |
A variation of draw.tree() from the maptree package.
[email protected], Denis White
Package home page: https://rattle.togaware.com
## this is usually used in the context of the plotRisk function ## Not run: drawTreeNodes(rpart(Species ~ ., iris))## this is usually used in the context of the plotRisk function ## Not run: drawTreeNodes(rpart(Species ~ ., iris))
Using the Rattle drawTreeNodes, draw a selection of Ada trees.
drawTreesAda(model, trees=0, title="")drawTreesAda(model, trees=0, title="")
model |
an ada model. |
trees |
The list of trees to draw. Use 0 to draw all trees. |
title |
An option title to add. |
Using Rattle's drawTreeNodes underneath, a plot for each of the specified trees from an Ada model will be displayed.
Package home page: https://rattle.togaware.com
## Not run: drawTreesAda(ds.ada)## Not run: drawTreesAda(ds.ada)
An error matrix reports the true/false potisitve/negative rates.
errorMatrix(actual, predicted, percentage=TRUE, digits=ifelse(percentage,1,3), count=FALSE)errorMatrix(actual, predicted, percentage=TRUE, digits=ifelse(percentage,1,3), count=FALSE)
actual |
a vector of true values. |
predicted |
a vector of predicted values. |
percentage |
return percentages. |
digits |
the number of digits to round results. |
count |
return counts. |
Package home page: https://rattle.togaware.com
## Not run: errorMatrix(model)## Not run: errorMatrix(model)
By taking predicted values, actual values, and measures of the risk associated with each case, generate a summary that groups the distinct predicted values, calculating the accumulative percentage Caseload, Recall, Risk, Precision, and Measure.
evaluateRisk(predicted, actual, risks)evaluateRisk(predicted, actual, risks)
predicted |
a numeric vector of probabilities (between 0 and 1) representing the probability of each entity being a 1. |
actual |
a numeric vector of classes (0 or 1). |
risks |
a numeric vector of risk (e.g., dollar amounts) associated with each entity that has a acutal of 1. |
Package home page: https://rattle.togaware.com
## simulate the data that is typical in data mining ## we often have only a small number of positive known case cases <- 1000 actual <- as.integer(rnorm(cases) > 1) adjusted <- sum(actual) nfa <- cases - adjusted ## risks might be dollar values associated adjusted cases risks <- rep(0, cases) risks[actual==1] <- round(abs(rnorm(adjusted, 10000, 5000)), 2) ## our models will generated a probability of a case being a 1 predicted <- rep(0.1, cases) predicted[actual==1] <- predicted[actual==1] + rnorm(adjusted, 0.3, 0.1) predicted[actual==0] <- predicted[actual==0] + rnorm(nfa, 0.1, 0.08) predicted <- signif(predicted) ## call upon evaluateRisk to generate performance summary ev <- evaluateRisk(predicted, actual, risks) ## have a look at the first few and last few head(ev) tail(ev) ## the performance is usually presented as a Risk Chart ## under the CRAN MS/Windows this causes a problem, so don't run for now ## Not run: plotRisk(ev$Caseload, ev$Precision, ev$Recall, ev$Risk)## simulate the data that is typical in data mining ## we often have only a small number of positive known case cases <- 1000 actual <- as.integer(rnorm(cases) > 1) adjusted <- sum(actual) nfa <- cases - adjusted ## risks might be dollar values associated adjusted cases risks <- rep(0, cases) risks[actual==1] <- round(abs(rnorm(adjusted, 10000, 5000)), 2) ## our models will generated a probability of a case being a 1 predicted <- rep(0.1, cases) predicted[actual==1] <- predicted[actual==1] + rnorm(adjusted, 0.3, 0.1) predicted[actual==0] <- predicted[actual==0] + rnorm(nfa, 0.1, 0.08) predicted <- signif(predicted) ## call upon evaluateRisk to generate performance summary ev <- evaluateRisk(predicted, actual, risks) ## have a look at the first few and last few head(ev) tail(ev) ## the performance is usually presented as a Risk Chart ## under the CRAN MS/Windows this causes a problem, so don't run for now ## Not run: plotRisk(ev$Caseload, ev$Precision, ev$Recall, ev$Risk)
Plots a fancy RPart decision tree using the pretty rpart plotter.
fancyRpartPlot(model, main="", sub, caption, palettes, type=2, ...)fancyRpartPlot(model, main="", sub, caption, palettes, type=2, ...)
model |
an rpart object. |
main |
title for the plot. |
sub |
sub title for the plot. The default is a Rattle string with date, time and username. |
caption |
caption for bottom right of plot. |
palettes |
a list of sequential palettes names. As supported by RColorBrewer::brewer.pal the available names are Blues BuGn BuPu GnBu Greens Greys Oranges OrRd PuBu PuBuGn PuRd Purples RdPu Reds YlGn YlGnBu YlOrBr YlOrRd. |
type |
the type of plot to generate (2). |
... |
additional arguments passed on to prp. |
Package home page: https://rattle.togaware.com
## Use rpart to build a decision tree. ## Not run: library(rpart) ## Set up the data for modelling. set.seed(42) ds <- weather target <- "RainTomorrow" risk <- "RISK_MM" ignore <- c("Date", "Location", risk) vars <- setdiff(names(ds), ignore) nobs <- nrow(ds) form <- formula(paste(target, "~ .")) train <- sample(nobs, 0.7*nobs) test <- setdiff(seq_len(nobs), train) actual <- ds[test, target] risks <- ds[test, risk] # Fit the model. fit <- rpart(form, data=ds[train, vars]) ## Plot the model. fancyRpartPlot(fit) ## Choose different colours. fancyRpartPlot(fit, palettes=c("Greys", "Oranges")) ## Add a main title to the plot. fancyRpartPlot(fit, main=target) ## End(Not run)## Use rpart to build a decision tree. ## Not run: library(rpart) ## Set up the data for modelling. set.seed(42) ds <- weather target <- "RainTomorrow" risk <- "RISK_MM" ignore <- c("Date", "Location", risk) vars <- setdiff(names(ds), ignore) nobs <- nrow(ds) form <- formula(paste(target, "~ .")) train <- sample(nobs, 0.7*nobs) test <- setdiff(seq_len(nobs), train) actual <- ds[test, target] risks <- ds[test, risk] # Fit the model. fit <- rpart(form, data=ds[train, vars]) ## Plot the model. fancyRpartPlot(fit) ## Choose different colours. fancyRpartPlot(fit, palettes=c("Greys", "Oranges")) ## Add a main title to the plot. fancyRpartPlot(fit, main=target) ## End(Not run)
Generate a string that is intended to be eval'd that
will add a title and sub-title to a plot. The string is a call to
title, supplying the given arguments,
pasted together, as the main title, and generating a
sub-title that begins with ‘Rattle’ and continues with the current date
and time, and finishes with the current user's username. This is used
internally in Rattle to adorn a plot with relevant information, but
may be useful outside of Rattle.
genPlotTitleCmd(..., vector=FALSE)genPlotTitleCmd(..., vector=FALSE)
... |
one or more strings that will be pasted together to form the main title. |
vector |
whether to return a vector as the result. |
Package home page: https://rattle.togaware.com
# generate some random plot plot(rnorm(100)) # generate the string representing the command to add titles tl <- genPlotTitleCmd("Sample Plot of", "No Particular Importance") # cause the string to be executed as an R command eval(parse(text=tl))# generate some random plot plot(rnorm(100)) # generate the string representing the command to add titles tl <- genPlotTitleCmd("Sample Plot of", "No Particular Importance") # cause the string to be executed as an R command eval(parse(text=tl))
Model.
ggVarImp(model, ...)ggVarImp(model, ...)
model |
object. |
... |
arguments passed on. |
Package home page: https://rattle.togaware.com
## Not run: ggVarImp(model)## Not run: ggVarImp(model)
Returns a list of the variables used and their frequencies.
listAdaVarsUsed(model)listAdaVarsUsed(model)
model |
an rpart object. |
Package home page: https://rattle.togaware.com
Display the textual representation of a selection of Ada trees.
listTreesAda(model, trees=0)listTreesAda(model, trees=0)
model |
an ada model. |
trees |
The list of trees to list. Use 0 to list all trees. |
Using rpart's print method display each of the specified trees from an Ada model.
Package home page: https://rattle.togaware.com
## Not run: listTreesAda(ds.ada)## Not run: listTreesAda(ds.ada)
Generate a list of packages installed and their version number.
listVersions(file="", ...)listVersions(file="", ...)
file |
a character string naming a file or a connection open for writing. '""' indicates output to the console. |
... |
arguments to |
This function is useful in reporting problems or bugs, to ensure there is a clear match of R package versions between the system exhibiting the issue and the test system replicating the issue.
By default the information is written to the console in a comma separated form, that is ideally designed to be written to a CSV file for emailing.
The mode is the most common or modal value of a list.
modalvalue(x, na.rm=FALSE)modalvalue(x, na.rm=FALSE)
x |
A vector, array or list. |
na.rm |
Whether to remove missing values. |
This function calculates the mode of a vector, array or list (lists are flattened). This code originated from an anonymous post on the R Wiki.
Plots a a vertical line at x up to max of y1 and y2, then horizontal from this line at y1 and y2. Intended for plotting on a plotRisk.
plotOptimalLine(x, y1, y2, pr = NULL, colour = "plum", label = NULL)plotOptimalLine(x, y1, y2, pr = NULL, colour = "plum", label = NULL)
x |
location of vertical line. |
y1 |
location of one horizontal line. |
y2 |
location of other horizontal line. |
pr |
Aprint a percentage at this point. |
colour |
of the line. |
label |
at bottom of line. |
Intended to plot an optimal line on a Risk Chart as plotted by plotRisk.
Package home page: https://rattle.togaware.com
## this is usually used in the context of the plotRisk function ## Not run: ev <- evaluateRisk(predicted, actual, risk) ## imitate this output here ev <- NULL ev$Caseload <- c(1.0, 0.8, 0.6, 0.4, 0.2, 0) ev$Precision <- c(0.15, 0.18, 0.21, 0.25, 0.28, 0.30) ev$Recall <- c(1.0, 0.95, 0.80, 0.75, 0.5, 0.0) ev$Risk <- c(1.0, 0.98, 0.90, 0.77, 0.30, 0.0) ## plot the Risk Chart plotRisk(ev$Caseload, ev$Precision, ev$Recall, ev$Risk, chosen=60, chosen.label="Pr=0.45") ## plot the optimal point plotOptimalLine(40, 77, 75, colour="maroon")## this is usually used in the context of the plotRisk function ## Not run: ev <- evaluateRisk(predicted, actual, risk) ## imitate this output here ev <- NULL ev$Caseload <- c(1.0, 0.8, 0.6, 0.4, 0.2, 0) ev$Precision <- c(0.15, 0.18, 0.21, 0.25, 0.28, 0.30) ev$Recall <- c(1.0, 0.95, 0.80, 0.75, 0.5, 0.0) ev$Risk <- c(1.0, 0.98, 0.90, 0.77, 0.30, 0.0) ## plot the Risk Chart plotRisk(ev$Caseload, ev$Precision, ev$Recall, ev$Risk, chosen=60, chosen.label="Pr=0.45") ## plot the optimal point plotOptimalLine(40, 77, 75, colour="maroon")
Plots a Rattle Risk Chart. Such a chart has been developed in a practical context to present the performance of data mining models to clients, plotting a caseload against performance, allowing a client to see the tradeoff between coverage and performance.
plotRisk(cl, pr, re, ri = NULL, title = NULL, show.legend = TRUE, xleg = 60, yleg = 55, optimal = NULL, optimal.label = "", chosen = NULL, chosen.label = "", include.baseline = TRUE, dev = "", filename = "", show.knots = NULL, show.lift=TRUE, show.precision=TRUE, risk.name = "Risk", recall.name = "Recall", precision.name = "Precision")plotRisk(cl, pr, re, ri = NULL, title = NULL, show.legend = TRUE, xleg = 60, yleg = 55, optimal = NULL, optimal.label = "", chosen = NULL, chosen.label = "", include.baseline = TRUE, dev = "", filename = "", show.knots = NULL, show.lift=TRUE, show.precision=TRUE, risk.name = "Risk", recall.name = "Recall", precision.name = "Precision")
cl |
a vector of caseloads corresponding to different probability cutoffs. Can be either percentages (between 0 and 100) or fractions (between 0 and 1). |
pr |
a vector of precision values for each probability cutoff. Can be either percentages (between 0 and 100) or fractions (between 0 and 1). |
re |
a vector of recall values for each probability cutoff. Can be either percentages (between 0 and 100) or fractions (between 0 and 1). |
ri |
a vector of risk values for each probability cutoff. Can be either percentages (between 0 and 100) or fractions (between 0 and 1). |
title |
the main title to place at the top of the plot. |
show.legend |
whether to display the legend in the plot. |
xleg |
the x coordinate for the placement of the legend. |
yleg |
the y coordinate for the placement of the legend. |
optimal |
a caseload (percentage or fraction) that represents an
optimal performance point which is also plotted. If instead the value
is |
optimal.label |
a string which is added to label the line drawn as the optimal point. |
chosen |
a caseload (percentage or fraction) that represents a user chosen optimal performance point which is also plotted. |
chosen.label |
a string which is added to label the line drawn as the chosen point. |
include.baseline |
if TRUE (the default) then display the diagonal baseline. |
dev |
a string which, if supplied, identifies a device type as
the target for the plot. This might be one of |
filename |
a string naming a file. If |
show.knots |
a vector of caseload values at which a vertical line should be drawn. These might correspond, for example, to individual paths through a decision tree, illustrating the impact of each path on the caseload and performance. |
show.lift |
whether to label the right axis with lift. |
show.precision |
whether to show the precision plot. |
risk.name |
a string used within the plot's legend that gives a
name to the risk. Often the risk is a dollar amount at risk from a
fraud or from a bank loan point of view, so the default is
|
recall.name |
a string used within the plot's legend that gives a
name to the recall. The recall is often the percentage of cases that
are positive hits, and in practise these might correspond to known
cases of fraud or reviews where some adjustment to perhaps a incom tax
return or application for credit had to be made on reviewing the case,
and so the default is |
precision.name |
a string used within the plot's legend that gives a
name to the precision. A common name for precision is |
Caseload is the percentage of the entities in the dataset covered by the model at a particular probability cutoff, so that with a cutoff of 0, all (100%) of the entities are covered by the model. With a cutoff of 1 (0%) no entities are covered by the model. A diagonal line is drawn to represent a baseline random performance. Then the percentage of positive cases (the recall) covered for a particular caseload is plotted, and optionally a measure of the percentage of the total risk that is also covered for a particular caseload may be plotted. Such a chart allows a user to select an appropriate tradeoff between caseload and performance. The charts are similar to ROC curves. The precision (i.e., strike rate) is also plotted.
Package home page: https://rattle.togaware.com
evaluateRisk, genPlotTitleCmd.
## this is usually used in the context of the evaluateRisk function ## Not run: ev <- evaluateRisk(predicted, actual, risk) ## imitate this output here ev <- NULL ev$Caseload <- c(1.0, 0.8, 0.6, 0.4, 0.2, 0) ev$Precision <- c(0.15, 0.18, 0.21, 0.25, 0.28, 0.30) ev$Recall <- c(1.0, 0.95, 0.80, 0.75, 0.5, 0.0) ev$Risk <- c(1.0, 0.98, 0.90, 0.77, 0.30, 0.0) ## plot the Risk Chart plotRisk(ev$Caseload, ev$Precision, ev$Recall, ev$Risk, chosen=60, chosen.label="Pr=0.45") ## Add a title eval(parse(text=genPlotTitleCmd("Sample Risk Chart")))## this is usually used in the context of the evaluateRisk function ## Not run: ev <- evaluateRisk(predicted, actual, risk) ## imitate this output here ev <- NULL ev$Caseload <- c(1.0, 0.8, 0.6, 0.4, 0.2, 0) ev$Precision <- c(0.15, 0.18, 0.21, 0.25, 0.28, 0.30) ev$Recall <- c(1.0, 0.95, 0.80, 0.75, 0.5, 0.0) ev$Risk <- c(1.0, 0.98, 0.90, 0.77, 0.30, 0.0) ## plot the Risk Chart plotRisk(ev$Caseload, ev$Precision, ev$Recall, ev$Risk, chosen=60, chosen.label="Pr=0.45") ## Add a title eval(parse(text=genPlotTitleCmd("Sample Risk Chart")))
A randomForest model, by default, consists of 500 decision trees. This function walks through each tree and generates a set of rules which are printed to the console. This takes a considerable amount of time and is provided for users to access the actual model, but it is not yet used within the Rattle GUI. It may be used to display the output of the RF (but it takes longer to generate than the model itself!). Or it might only be used on export to PMML or SQL.
printRandomForests(model, models=NULL, include.class=NULL, format="")printRandomForests(model, models=NULL, include.class=NULL, format="")
model |
a randomForest model. |
models |
a list of integers limiting the models in MODEL that are displayed. |
include.class |
limit the output to the specific class. |
format |
possible values are "VB". |
Package home page: https://rattle.togaware.com
## Display a ruleset for a specific model amongst the 500. ## Not run: printRandomForests(rfmodel, 5) ## Display a ruleset for specific models amongst the 500. ## Not run: printRandomForests(rfmodel, c(5,10,15)) ## Display a ruleset for each of the 500 models. ## Not run: printRandomForests(rfmodel)## Display a ruleset for a specific model amongst the 500. ## Not run: printRandomForests(rfmodel, 5) ## Display a ruleset for specific models amongst the 500. ## Not run: printRandomForests(rfmodel, c(5,10,15)) ## Display a ruleset for each of the 500 models. ## Not run: printRandomForests(rfmodel)
A randomForest model, by default, consists of 500 decision trees. This function walks through each tree and generates a set of rules. This takes a considerable amount of time and is provided for users to access the actual model, but it is not yet used within the Rattle GUI. It may be used to display the output of the RF (but it takes longer to generate than the model itself!). Or it might only be used on export to PMML or SQL.
randomForest2Rules(model, models=NULL)randomForest2Rules(model, models=NULL)
model |
a randomForest model. |
models |
a list of integers limiting the models in MODEL that are converted. |
Package home page: https://rattle.togaware.com
## Generate a ruleset for a specific model amongst the 500. ## Not run: randomForest2Rules(rfmodel, 5) ## Generate a ruleset for specific models amongst the 500. ## Not run: randomForest2Rules(rfmodel, c(5,10,15)) ## Generate a ruleset for each of the 500 models. ## Not run: randomForest2Rules(rfmodel)## Generate a ruleset for a specific model amongst the 500. ## Not run: randomForest2Rules(rfmodel, 5) ## Generate a ruleset for specific models amongst the 500. ## Not run: randomForest2Rules(rfmodel, c(5,10,15)) ## Generate a ruleset for each of the 500 models. ## Not run: randomForest2Rules(rfmodel)
The Rattle user interface uses the RGtk2 package to present an intuitive point and click interface for data mining, extensively building on the excellent collection of R packages by very many authors for data manipulation, exploration, analysis, and evaluation.
rattle(csvname=NULL, dataset=NULL, useGtkBuilder=TRUE)rattle(csvname=NULL, dataset=NULL, useGtkBuilder=TRUE)
csvname |
the optional name of a CSV file to load into Rattle on startup. |
dataset |
The optional name as a character string of a dataset to load into Rattle on startup. |
useGtkBuilder |
if not supplied then automatically determine whether to use the new GtkBuilder rather than the deprecated libglade. A user can override the heuristic choice with TRUE or FALSE. |
Refer to the Rattle home page in the URL below for a growing reference manual for using Rattle.
Whilst the underlying functionality of Rattle is built upon a vast collection of other R packages, Rattle itself provides a collection of utility functions used within Rattle. These are made available through loading the rattle package into your R library. The See Also section lists these utility functions that may be useful outside of Rattle.
Rattle can initialise some options using a .Rattle file if the folder in which Rattle is started. The currently supported options are .RATTLE.DATA, .RATTLE.SCORE.IN, and .RATTLE.SCORE.OUT.
If the environment variable RATTLE\_DATA is defined then that is set as the default CSV file name to load. Otherwise, if .RATTLE.DATA is defined then that will be used as the CSV file to load. Otherwise, if csvname is provided then that will be used.
Two environments are exported by Rattle, capturing the current rattle state (crs) and the current rattle variables (crv).
Package home page: https://rattle.togaware.com
evaluateRisk, genPlotTitleCmd,
plotRisk.
# You can start rattle with a path to a csv file to pre-specify the # dataset. You then need to click Execute to load the data. ## Not run: rattle(system.file("csv", "weather.csv", package = "rattle"))# You can start rattle with a path to a csv file to pre-specify the # dataset. You then need to click Execute to load the data. ## Not run: rattle(system.file("csv", "weather.csv", package = "rattle"))
Displays a textual reveiw of the performance of a multinom model.
rattle.print.summary.multinom(x, digits = x$digits, ...)rattle.print.summary.multinom(x, digits = x$digits, ...)
x |
An rpart object. |
digits |
Number of digist to print for numbers. |
... |
Other arguments. |
Print a summary of a multinom model. This is sipmly a modification of the print.summary.multinom function to add the number of entities!
Package home page: https://rattle.togaware.com
Display system information, including versions of Rattle and R, operating system, and versions of other packages used by Rattle. Useful for reporting bugs but also invisibly returns a list of packages that have updates available and can be passed to install.packages().
rattleInfo(all.dependencies=FALSE, include.not.installed=FALSE, include.not.available=FALSE, include.libpath=FALSE)rattleInfo(all.dependencies=FALSE, include.not.installed=FALSE, include.not.available=FALSE, include.libpath=FALSE)
all.dependencies |
If TRUE then check the full dependency graph for Rattle and list all of those packages (which may take quite a few seconds to compute), or else just list those key packages that Rattle Depends on and Suggests. |
include.not.installed |
If TRUE then make mention of any packages that are not installed, but are available. |
include.not.available |
If TRUE then make mention of any packages that are not available from CRAN. |
include.libpath |
If TRUE then list the library location where each package is installed. |
This is a support function to list useful information to provide the developers with information about the system environment when running Rattle. It is intended to provide the information that is useful in reporting bugs.
It also lists the currently installed version of a number of packages that Rattle makes use of as well as checking for any updates available for those packages.
If updates are found then a command is generated and printed so that a user can simply copy and paste the command to update the relevant packages. The function also invisibly returns the list of packages that can be updated, so that we can do something like: install.packages(rattleInfo()).
Package home page: https://rattle.togaware.com
The numeric vector is remapped to integers from 0 to max-1, with any missing values mapped to the midpoint. Original idea from Tony Nolan. This will eventually be generalised to do the remapping using any of the rescaling functions.
rescale.by.group(x, by=NULL, type = "irank", itop = 100)rescale.by.group(x, by=NULL, type = "irank", itop = 100)
x |
The numeric vector to rescale. |
by |
A factor of the same length as x used to define the groups. |
type |
The type of rescaling to perform. |
itop |
For an integer remapping this is the number of groups, so that the numeric values are maped to the integers from 0 to (max-1). |
This Rattle support function, which is also useful by itself, provides a simple mechanism to rescale a numeric variable. Several rescalings are possible. The rescaling is done by first grouping the observations according to the by argument.
Package home page: https://rattle.togaware.com
Plots a Rattle Risk Chart for binary classification models using ggplot2. Such a chart has been developed in a practical context to present the performance of data mining models to clients, plotting a caseload against performance, allowing a client to see the tradeoff between coverage and performance.
riskchart(pr, ac, ri = NULL, title = "Risk Chart", title.size = 10, subtitle = NULL, caption = TRUE, show.legend = TRUE, optimal = NULL, optimal.label = "", chosen = NULL, chosen.label = "", include.baseline = TRUE, dev = "", filename = "", show.knots = NULL, show.lift = TRUE, show.precision = TRUE, show.maximal = TRUE, risk.name = "Risk", recall.name = "Recall", precision.name = "Precision", thresholds = NULL, legend.horiz = TRUE)riskchart(pr, ac, ri = NULL, title = "Risk Chart", title.size = 10, subtitle = NULL, caption = TRUE, show.legend = TRUE, optimal = NULL, optimal.label = "", chosen = NULL, chosen.label = "", include.baseline = TRUE, dev = "", filename = "", show.knots = NULL, show.lift = TRUE, show.precision = TRUE, show.maximal = TRUE, risk.name = "Risk", recall.name = "Recall", precision.name = "Precision", thresholds = NULL, legend.horiz = TRUE)
pr |
The predicted class for each observation. |
ac |
The actual class for each observation. |
ri |
The risk class for each observation. |
title |
the main title to place at the top of the plot. |
title.size |
font size for the main title. |
subtitle |
subtitle under the main title. |
caption |
caption for the bottom right of plot. |
show.legend |
whether to display the legend in the plot. |
optimal |
a caseload (percentage or fraction) that represents an
optimal performance point which is also plotted. If instead the value
is |
optimal.label |
a string which is added to label the line drawn as the optimal point. |
chosen |
a caseload (percentage or fraction) that represents a user chosen optimal performance point which is also plotted. |
chosen.label |
a string which is added to label the line drawn as the chosen point. |
include.baseline |
if TRUE (the default) then display the diagonal baseline. |
dev |
a string which, if supplied, identifies a device type as
the target for the plot. This might be one of |
filename |
a string naming a file. If |
show.knots |
a vector of caseload values at which a vertical line should be drawn. These might correspond, for example, to individual paths through a decision tree, illustrating the impact of each path on the caseload and performance. |
show.lift |
whether to label the right axis with lift. |
show.precision |
whether to show the precision plot. |
show.maximal |
whether to show the maximal performance line. |
risk.name |
a string used within the plot's legend that gives a
name to the risk. Often the risk is a dollar amount at risk from a
fraud or from a bank loan point of view, so the default is
|
recall.name |
a string used within the plot's legend that gives a
name to the recall. The recall is often the percentage of cases that
are positive hits, and in practise these might correspond to known
cases of fraud or reviews where some adjustment to perhaps a incom tax
return or application for credit had to be made on reviewing the case,
and so the default is |
precision.name |
a string used within the plot's legend that gives
a name to the precision. A common name for precision is |
thresholds |
whether to display scores along the top axis. |
legend.horiz |
whether to display a horizontal legend. |
Caseload is the percentage of the entities in the dataset covered by the model at a particular probability cutoff, so that with a cutoff of 0, all (100%) of the entities are covered by the model. With a cutoff of 1 (0%) no entities are covered by the model. A diagonal line is drawn to represent a baseline random performance. Then the percentage of positive cases (the recall) covered for a particular caseload is plotted, and optionally a measure of the percentage of the total risk that is also covered for a particular caseload may be plotted. Such a chart allows a user to select an appropriate tradeoff between caseload and performance. The charts are similar to ROC curves. The precision (i.e., strike rate) is also plotted.
Package home page: https://rattle.togaware.com
evaluateRisk, genPlotTitleCmd.
## Not run: ## Use rpart to build a decision tree. library(rpart) ## Set up the data for modelling. set.seed(42) ds <- weather target <- "RainTomorrow" risk <- "RISK_MM" ignore <- c("Date", "Location", risk) vars <- setdiff(names(ds), ignore) nobs <- nrow(ds) form <- formula(paste(target, "~ .")) train <- sample(nobs, 0.7*nobs) test <- setdiff(seq_len(nobs), train) actual <- ds[test, target] risks <- ds[test, risk] # Build the model. model <- rpart(form, data=ds[train, vars]) ## Obtain predictions. predicted <- predict(model, ds[test, vars], type="prob")[,2] ## Plot the Risk Chart. riskchart(predicted, actual, risks) ## End(Not run)## Not run: ## Use rpart to build a decision tree. library(rpart) ## Set up the data for modelling. set.seed(42) ds <- weather target <- "RainTomorrow" risk <- "RISK_MM" ignore <- c("Date", "Location", risk) vars <- setdiff(names(ds), ignore) nobs <- nrow(ds) form <- formula(paste(target, "~ .")) train <- sample(nobs, 0.7*nobs) test <- setdiff(seq_len(nobs), train) actual <- ds[test, target] risks <- ds[test, risk] # Build the model. model <- rpart(form, data=ds[train, vars]) ## Obtain predictions. predicted <- predict(model, ds[test, vars], type="prob")[,2] ## Plot the Risk Chart. riskchart(predicted, actual, risks) ## End(Not run)
For the current device, or for the device identified, save the plot displayed there in some way. This is either saved to file, copied to the clipboard for pasting into other applications, or sent to the printer for saving a hard copy.
savePlotToFile(file.name, dev.num=dev.cur()) copyPlotToClipboard(dev.num=dev.cur()) printPlot(dev.num=dev.cur())savePlotToFile(file.name, dev.num=dev.cur()) copyPlotToClipboard(dev.num=dev.cur()) printPlot(dev.num=dev.cur())
file.name |
Character string naming the file including the file name extension which is used to specify the type of file to save. |
dev.num |
A device number indicating which device to save. |
Package home page: https://rattle.togaware.com
This rattle support function is used for encapsulating data mining objects. The supplied environment is augmented with other data derived from the supplied data, such as a sample trianing dataset, list of numeric variables, and a formula for modelling.
setupDataset(env, seed=NULL)setupDataset(env, seed=NULL)
env |
the environment to modify. |
seed |
optionally set the seed for repeatability. |
The supplied object (an environment) is assumed to also contain the variables data (a data frame), target (a character string naming the target variable), risk (a character string naming the risk variable), and inputs (a character vector naming all the input variables). This function then adds in the variables vars (the variables used for modelling), numerics (the numeric vars within inputs), nobs (the number of observations), form (the formula for building models), train (a 70% training dataset).
Package home page: https://rattle.togaware.com
Often we want to view the actual trees built by a random forest. Although reviewing all 500 trees might be a bit much, this function allows us to at least list them.
treeset.randomForest(model, n=1, root=1, format="R")treeset.randomForest(model, n=1, root=1, format="R")
model |
a randomForest model. |
n |
a specific tree to list. |
root |
where to start the stree from, primarily for internal use. |
format |
one of "R", "VB". |
Package home page: https://rattle.togaware.com
## Display a treeset for a specific model amongst the 500. ## Not run: treeset.randomForests(rfmodel, 5)## Display a treeset for a specific model amongst the 500. ## Not run: treeset.randomForests(rfmodel, 5)
One year of daily weather observations collected from the Canberra airport in Australia was obtained from the Australian Commonwealth Bureau of Meteorology and processed to create this sample dataset for illustrating data mining using R and Rattle.
The data has been processed to provide a target variable
RainTomorrow (whether there is rain on the following day -
No/Yes) and a risk variable RISK_MM (how much rain recorded in
millimetres). Various transformations were performed on the source
data. The dataset is quite small and is useful only for repeatable
demonstration of various data science operations.
The source dataset is Copyright by the Australian Commonwealth Bureau of Meteorology and is provided as part of the rattle package with permission.
weatherweather
The weather dataset is a data frame containing one year of
daily observations from a single weather station (Canberra).
DateThe date of observation (a Date object).
LocationThe common name of the location of the weather station.
MinTempThe minimum temperature in degrees celsius.
MaxTempThe maximum temperature in degrees celsius.
RainfallThe amount of rainfall recorded for the day in mm.
EvaporationThe so-called Class A pan evaporation (mm) in the 24 hours to 9am.
SunshineThe number of hours of bright sunshine in the day.
WindGustDirThe direction of the strongest wind gust in the 24 hours to midnight.
WindGustSpeedThe speed (km/h) of the strongest wind gust in the 24 hours to midnight.
Temp9amTemperature (degrees C) at 9am.
RelHumid9amRelative humidity (percent) at 9am.
Cloud9amFraction of sky obscured by cloud at 9am. This is measured in "oktas", which are a unit of eigths. It records how many eigths of the sky are obscured by cloud. A 0 measure indicates completely clear sky whilst an 8 indicates that it is completely overcast.
WindSpeed9amWind speed (km/hr) averaged over 10 minutes prior to 9am.
Pressure9amAtmospheric pressure (hpa) reduced to mean sea level at 9am.
Temp3pmTemperature (degrees C) at 3pm.
RelHumid3pmRelative humidity (percent) at 3pm.
Cloud3pmFraction of sky obscured by cloud (in "oktas": eighths) at 3pm. See Cload9am for a description of the values.
WindSpeed3pmWind speed (km/hr) averaged over 10 minutes prior to 3pm.
Pressure3pmAtmospheric pressure (hpa) reduced to mean sea level at 3pm.
ChangeTempChange in temperature.
ChangeTempDirDirection of change in temperature.
ChangeTempMagMagnitude of change in temperature.
ChangeWindDirectDirection of wind change.
MaxWindPeriodPeriod of maximum wind.
RainTodayInteger: 1 if precipitation (mm) in the 24 hours to 9am exceeds 1mm, otherwise 0.
TempRangeDifference between minimum and maximum temperatures (degrees C) in the 24 hours to 9am.
PressureChangeChange in pressure.
RISK_MMThe amount of rain. A kind of measure of the "risk".
RainTomorrowThe target variable. Did it rain tomorrow?
The daily observations are available from https://www.bom.gov.au/climate/data. Copyright Commonwealth of Australia 2010, Bureau of Meteorology.
Definitions adapted from https://www.bom.gov.au/climate/dwo/IDCJDW0000.shtml
Package home page: https://rattle.togaware.com. Data source: https://www.bom.gov.au/climate/dwo/ and https://www.bom.gov.au/climate/data.
Daily weather observations from multiple locations around Australia, obtained from the Australian Commonwealth Bureau of Meteorology and processed to create this realtively large sample dataset for illustrating analytics, data mining, and data science using R and Rattle.
The data has been processed to provide a target variable
RainTomorrow (whether there is rain on the following day -
No/Yes) and a risk variable RISK_MM (how much rain recorded in
millimeters). Various transformations are performed on the data.
The weatherAUS dataset is regularly updated an updates of this
package usually correspond to updates to this dataset. The data is
updated from the Bureau of Meteorology web site.
The locationsAUS dataset records the location of each weather
station.
The source dataset comes from the Australian Commonwealth Bureau of Meteorology. The Bureau provided permission to use the data with the Bureau of Meteorology acknowledged as the source of the data, as per email from Cathy Toby ([email protected]) of the Climate Information Services of the National CLimate Centre, 17 Dec 2008.
A CSV version of this dataset is available as https://rattle.togaware.com/weatherAUS.csv.
weatherAUSweatherAUS
The weatherAUS dataset is a data frame containing over 140,000
daily observations from over 45 Australian weather stations.
DateThe date of observation (a Date object).
LocationThe common name of the location of the weather station.
MinTempThe minimum temperature in degrees celsius.
MaxTempThe maximum temperature in degrees celsius.
RainfallThe amount of rainfall recorded for the day in mm.
EvaporationThe so-called Class A pan evaporation (mm) in the 24 hours to 9am.
SunshineThe number of hours of bright sunshine in the day.
WindGustDirThe direction of the strongest wind gust in the 24 hours to midnight.
WindGustSpeedThe speed (km/h) of the strongest wind gust in the 24 hours to midnight.
Temp9amTemperature (degrees C) at 9am.
RelHumid9amRelative humidity (percent) at 9am.
Cloud9amFraction of sky obscured by cloud at 9am. This is measured in "oktas", which are a unit of eigths. It records how many eigths of the sky are obscured by cloud. A 0 measure indicates completely clear sky whilst an 8 indicates that it is completely overcast.
WindSpeed9amWind speed (km/hr) averaged over 10 minutes prior to 9am.
Pressure9amAtmospheric pressure (hpa) reduced to mean sea level at 9am.
Temp3pmTemperature (degrees C) at 3pm.
RelHumid3pmRelative humidity (percent) at 3pm.
Cloud3pmFraction of sky obscured by cloud (in "oktas": eighths) at 3pm. See Cload9am for a description of the values.
WindSpeed3pmWind speed (km/hr) averaged over 10 minutes prior to 3pm.
Pressure3pmAtmospheric pressure (hpa) reduced to mean sea level at 3pm.
ChangeTempChange in temperature.
ChangeTempDirDirection of change in temperature.
ChangeTempMagMagnitude of change in temperature.
ChangeWindDirectDirection of wind change.
MaxWindPeriodPeriod of maximum wind.
RainTodayInteger: 1 if precipitation (mm) in the 24 hours to 9am exceeds 1mm, otherwise 0.
TempRangeDifference between minimum and maximum temperatures (degrees C) in the 24 hours to 9am.
PressureChangeChange in pressure.
RISK_MMThe amount of rain. A kind of measure of the "risk".
RainTomorrowThe target variable. Did it rain tomorrow?
Observations were drawn from numerous weather stations. The daily observations are available from https://www.bom.gov.au/climate/data. Copyright Commonwealth of Australia 2010, Bureau of Meteorology.
Definitions adapted from https://www.bom.gov.au/climate/dwo/IDCJDW0000.shtml
Package home page: https://rattle.togaware.com. Data source: https://www.bom.gov.au/climate/dwo/ and https://www.bom.gov.au/climate/data.
A rattle support function.
whichNumerics(data)whichNumerics(data)
data |
a data frame. |
Package home page: https://rattle.togaware.com
The wine dataset contains the results of a chemical analysis of
wines grown in a specific area of Italy. Three types of wine are
represented in the 178 samples, with the results of 13 chemical
analyses recorded for each sample. The Type variable has been
transformed into a categoric variable.
The data contains no missing values and consits of only numeric data,
with a three class target variable (Type) for classification.
winewine
A data frame containing 178 observations of 13 variables.
TypeThe type of wine, into one of three classes, 1 (59 obs), 2(71 obs), and 3 (48 obs).
AlcoholAlcohol
MalicMalic acid
AshAsh
AlcalinityAlcalinity of ash
MagnesiumMagnesium
PhenolsTotal phenols
FlavanoidsFlavanoids
NonflavanoidsNonflavanoid phenols
ProanthocyaninsProanthocyanins
ColorColor intensity.
HueHue
DilutionD280/OD315 of diluted wines.
ProlineProline
The data was downloaded from the UCI Machine Learning Repository.
It was read as a CSV file with no header using
read.csv. The columns were then given the appropriate
names using colnames and the Type was transformed into a
factor using as.factor. The compressed R data file was
saved using save:
UCI <- "https://archive.ics.uci.edu/ml"
REPOS <- "machine-learning-databases"
wine.url <- sprintf("
wine <- read.csv(wine.url, header=FALSE)
colnames(wine) <- c('Type', 'Alcohol', 'Malic', 'Ash',
'Alcalinity', 'Magnesium', 'Phenols',
'Flavanoids', 'Nonflavanoids',
'Proanthocyanins', 'Color', 'Hue',
'Dilution', 'Proline')
wine$Type <- as.factor(wine$Type)
save(wine, file="wine.Rdata", compress=TRUE)
Asuncion, A. & Newman, D.J. (2007). UCI Machine Learning Repository [https://www.ics.uci.edu/~mlearn/MLRepository.html]. Irvine, CA: University of California, School of Information and Computer Science.