| Title: | Classes and Methods for Cross Validation of "Class Prediction" Algorithms |
|---|---|
| Description: | Defines classes and methods to cross-validate various binary classification algorithms used for "class prediction" problems. |
| Authors: | Kevin R. Coombes |
| Maintainer: | Kevin R. Coombes <[email protected]> |
| License: | Apache License (== 2.0) |
| Version: | 2.3.4 |
| Built: | 2024-08-30 06:25:29 UTC |
| Source: | CRAN |
When performing cross-validation on a dataset, it often becomes necessary to split the data into training and test sets that are balanced for a factor. This function implements such a balanced split.
balancedSplit(fac, size)balancedSplit(fac, size)
fac |
A factor that should be balanced between the two subsets. |
size |
A number between 0 and 1 indicating the fraction of the dataset to be used for training. |
This function randomly samples the same fraction of items from each level of a factor to include in a training set. In most cases, this will be a binary factor (and might even be the outcome that one wants to predict). However, the implementation works for factors with an arbitrary number of levels.
Returns a logical vector with length equal to the length of
fac. TRUE values designate samples selected for the training
set.
Kevin R. Coombes <[email protected]>
CrossValidate, CrossValidate-class.
nFeatures <- 40 nSamples <- 2*10 dataset <- matrix(rnorm(nSamples*nFeatures), ncol=nSamples) groups <- factor(rep(c("A", "B"), each=10)) balancedSplit(dataset, groups)nFeatures <- 40 nSamples <- 2*10 dataset <- matrix(rnorm(nSamples*nFeatures), ncol=nSamples) groups <- factor(rep(c("A", "B"), each=10)) balancedSplit(dataset, groups)
Given a model classifier and a data set, this function performs cross-validation by repeatedly splitting the data into training and testing subsets in order to estimate the performance of this kind of classifer on new data.
CrossValidate(model, data, status, frac, nLoop, prune=keepAll, verbose=TRUE)CrossValidate(model, data, status, frac, nLoop, prune=keepAll, verbose=TRUE)
model |
An element of the |
data |
A matrix containing the data to be used for cross-validation. As with most gene expression data, columns are the independent samples or observations and rows are the measured features. |
status |
A binary-valued factor with the classes to be predicted. |
frac |
A number between 0 and 1; the fraction of the data that should be used in each iteration to train the model. |
nLoop |
An integer; the number of times to split the data into training and test sets. |
prune |
A function that takes two inoputs, a data matrix and a factor with two levels, and rteturns a logical vector whose length equals the number of rows in the data matrix. |
verbose |
A logical value; should the cross-validation routine report interim progress. |
The CrossValidate package provides generic tools for performing
cross-validation on classificaiton methods in the context of
high-throughput data sets such as those produced by gene expression
microarrays. In order to use a classifier with this implementaiton of
cross-validation, you must first prepare a pair of functions (one for
learning models from training data, and one for making predictions on
test data). These functions, along with any required meta-parameters,
are used to create an object of the Modeler-class. That
object is then passed to the CrossValidate function along
with the full training data set. The full data set is then repeatedly
split into its own training and test sets; you can specify the
fraction to be used for training and the number of iterations. The
result is a detailed look at the accuracy, sensitivity, specificity,
and positive and negative predictive value of the model, as estimated
by cross-validation.
An object of the CrossValidate-class.
Kevin R. Coombes [email protected]
See the manual page for the CrossValidate-class for a list
of related references.
See CrossValidate-class for a description of the slots in
the object created by this function.
dataset <- matrix(rnorm(50*100), nrow=50) pseudoclass <- factor(rep(c("A", "B"), each=50)) model <- modelerCCP # obviously, other models can be used numTimes <- 10 # and more is probably better cv <- CrossValidate(model, dataset, pseudoclass, 0.5, numTimes) summary(cv)dataset <- matrix(rnorm(50*100), nrow=50) pseudoclass <- factor(rep(c("A", "B"), each=50)) model <- modelerCCP # obviously, other models can be used numTimes <- 10 # and more is probably better cv <- CrossValidate(model, dataset, pseudoclass, 0.5, numTimes) summary(cv)
A class that contains the results of internal cross-validation (by multiple splits into training and test sets) of an algorithm that builds a model to predict a binary outcome.
Objects should be created by calls to the constructor function,
CrossValidate.
nIterations:An integer; the number of times the data was split into training and test sets.
trainPercent:A number between 0 and 1; the fraction of data used in each training set.
outcome:A binary factor containing the true outcome for each sample.
trainOutcome:A data frame containing the true outcomes for each member of the training set. The value 'NA' is used for samples that were reserved for testing. Each column is a different split into training and test sets.
trainPredict:A data frame containing the predicted outcome from the model for each member of the training set. The value 'NA' is used for samples that were reserved for testing. Each column is a different split into training and test sets.
validOutcome:A data frame containing the true outcomes for each member of the test set. The value 'NA' is used for samples that were used for training. Each column is a different split into training and test sets.
validPredict:A data frame containing the predicted outcome from the model for each member of the test set. The value 'NA' is used for samples that were used for training. Each column is a different split into training and test sets.
extras:A list, whose length equals the number of plsits into trainin and test sets. Each entry contains any "extra" information collected during the fitting of the model; the kinds of items stored here depend on the actual classification algorithm used.
signature(object = "CrossValidate"): Produces
a summary of the performance of the algorithm on both the trinaing
sets and the test sets, in terms of specificity, sensitivity, and
positive or negative predictive value. Specifically, this method
returns an object of the CrossValSummary-class.
Kevin R. Coombes <[email protected]>
Braga-Neto U, Dougherty ER.
Is cross-validation valid for small-sample microarray
classification?
Bioinformatics, 2004; 20:374–380.
Jiang W, Varma S, Simon R.
Calculating confidence intervals for
prediction error in microarray classification using resampling.
Stat Appl Genet Mol Biol. 2008; 7:Article8.
Fu LM, Youn ES.
Improving reliability of gene selection from
microarray functional genomics data.
IEEE Trans Inf Technol Biomed. 2003; 7:191–6.
Man MZ, Dyson G, Johnson K, Liao B.
Evaluating methods for classifying expression data.
J Biopharm Stat. 2004; 14:1065–84.
Fu WJ, Carroll RJ, Wang S.
Estimating misclassification error with small samples via
bootstrap cross-validation.
Bioinformatics, 2005; 21:1979–86.
Ancona N, Maglietta R, Piepoli A, D'Addabbo A, Cotugno R, Savino M,
Liuni S, Carella M, Pesole G, Perri F.
On the statistical assessment of classifiers using DNA
microarray data.
BMC Bioinformatics, 2006; 7:387.
Lecocke M, Hess K.
An empirical study of univariate and genetic
algorithm-based feature selection in binary classification with
microarray data.
Cancer Inform, 2007; 2:313–27.
Lee S.
Mistakes in validating the accuracy of a prediction classifier
in high-dimensional but small-sample microarray data.
Stat Methods Med Res, 2008; 17:635–42.
See CrossValidate for the constructor function.
showClass("CrossValidate")showClass("CrossValidate")
Represents the effect of summarizing a CrossValidate-class object
by computing the performance of predictions on each split into
training and test sets.
Objects are almost always created automatically by applying the
summary method to an object of the
CrossValidate-class.
call:A "call" object recoding how the summary
method was invoked.
parent:A character vector containing the name of the
CrossValidate object being summarized.
trainAcc:A "list" containing five numeric
vectors: the sensitivity, specificity, accuracy, PPV, and NPV on
each training data set.
validAcc:A "list" containing five numeric
vectors: the sensitivity, specificity, accuracy, PPV, and NPV on
each test data set.
signature(object = "CrossValSummary"): Summarizes
the algorithm performance.
Kevin R. Coombes <[email protected]>
showClass("CrossValSummary")showClass("CrossValSummary")