kTSPclassifier: Classification Using k Pairs of Features With Relative...

In ClassifyR: A framework for cross-validated classification problems, with applications to differential variability and differential distribution testing

Description

Each pair of features votes for a class based on whether the value of one feature is less than the other feature.

Usage

  ## S4 method for signature 'matrix'
kTSPclassifier(measurements, classes, test, featurePairs = NULL, ...)
  ## S4 method for signature 'DataFrame'
kTSPclassifier(measurements, classes, test, featurePairs = NULL,
                   weighted = c("unweighted", "weighted", "both"),
                   minDifference = 0, returnType = c("class", "score", "both"), verbose = 3)
  ## S4 method for signature 'MultiAssayExperiment'
kTSPclassifier(measurements, test, target = names(measurements)[1],
                   featurePairs = NULL, ...)

Arguments

measurements	Either a matrix, DataFrame or MultiAssayExperiment containing the training data. For a matrix, the rows are features, and the columns are samples.
classes	Either a vector of class labels of class factor of the same length as the number of samples in measurements or if the measurements are of class DataFrame a character vector of length 1 containing the column name in measurement is also permitted. Not used if measurements is a MultiAssayExperiment object.
test	An object of the same class as measurements with no samples in common with measurements and the same number of features as it.
featurePairs	An object of class as Pairs containing the pairs of features to determine whether the inequality of the first feature being less than the second feature holds, indiciating evidence for the second level of the classes factor.
target	If measurements is a MultiAssayExperiment, the name of the data table to be used. "clinical" is also a valid value and specifies that integer variables from the clinical data table will be used.
...	Unused variables by the methods for a matrix or a MultiAssayExperiment passed to the DataFrame method which does the classification.
weighted	Default: "both". Either "both", "unweighted" or "weighted". In weighted mode, the difference in densities is summed over all features. If unweighted mode, each feature's vote is worth the same. Both can be calculated simultaneously.
minDifference	Default: 0. The minimum difference in densities for a feature to be allowed to vote. Can be a vector of cutoffs. If no features for a particular sample have a difference large enough, the class predicted is simply the largest class.
returnType	Default: "class". Either "class", "score" or "both". Sets the return value from the prediction to either a vector of class labels, score for a sample belonging to the second class, as determined by the factor levels, or both labels and scores in a data.frame.
verbose	Default: 3. A number between 0 and 3 for the amount of progress messages to give. This function only prints progress messages if the value is 3.

Details

If weighted is TRUE, then a sample's predicted class is based on the sum of differences of measurements for each feature pair. Otherwise, when weighted is FALSE, each pair of features has an equal vote, the predicted class is the one with the most votes. If the voting is tied, the the class with the most samples in the training set is voted for.

Because this method compares different features, they need to have comparable measurements. For example, RNA-seq counts would be unsuitable since these depend on the length of a feature, whereas F.P.K.M. values would be suitable.

The featurePairs to use is recommended to be determined in conjunction with pairsDifferencesSelection.

Value

A vector or list of class prediction information, as long as the number of samples in the test data, or lists of such information, if a variety of predictions is generated.

Author(s)

Dario Strbenac

See Also

pairsDifferencesSelection for a function which could be used to do feature selection before the k-TSP classifier is run.

Examples

  # Difference in differences for features A and C between classes.                                           
  measurements <- matrix(c(9.9, 10.5, 10.1, 10.9, 11.0, 6.6, 7.7, 7.0, 8.1, 6.5,
                           8.5, 10.5, 12.5, 10.5, 9.5, 8.5, 10.5, 12.5, 10.5, 9.5,
                           6.6, 7.7, 7.0, 8.1, 6.5, 11.2, 11.0, 11.1, 11.4, 12.0,
                           8.1, 10.6, 7.4, 7.1, 10.4, 6.1, 7.3, 2.7, 11.0, 9.1,
                           round(rnorm(60, 8, 1), 1)), ncol = 10, byrow = TRUE)
  classes <- factor(rep(c("Good", "Poor"), each = 5))
                         
  rownames(measurements) <- LETTERS[1:10]
  colnames(measurements) <- names(classes) <- paste("Patient", 1:10)
  
  trainIndex <- c(1:4, 6:9)
  trainMatrix <- measurements[, trainIndex]
  testMatrix <- measurements[, c(5, 10)]
  
  featurePairs <- Pairs('A', 'C') # Could be selected by pairsDifferencesSelection function.
  kTSPclassifier(trainMatrix, classes[trainIndex], testMatrix, featurePairs)

ClassifyR documentation built on Nov. 8, 2020, 6:53 p.m.