train_one_vs_rest_TSP {multiclassPairs}

Build multiclass rule-based classifier as one-vs-rest scheme

Description

train_one_vs_rest_TSP trains multiclass classifier in a one-vs-rest scheme by combining binary classifiers for each class produced by switchBox package.

Usage

train_one_vs_rest_TSP(data_object,
                      filtered_genes,
                      k_range = 10:50,
                      include_pivot = FALSE,
                      one_vs_one_scores = FALSE,
                      platform_wise_scores = FALSE,
                      disjoint = TRUE,
                      seed = NULL,
                      classes,
                      SB_arg = list(),
                      verbose = TRUE)

Arguments

data_object

data object generated by ReadData function. Object contains the data and labels.

filtered_genes

filtered genes object produced by filter_genes_TSP function

k_range

an integer or range represent the candidate number of Top Scoring Pairs (TSPs) in the individual (i.e. binary) classifiers. Default range from 10 to 50.

include_pivot

a logical indicating if the filtered genes should also be paired with all features available in the data matrix. Default is FALSE. include_pivot=FALSE means filtered genes will be paired with themselves only.

one_vs_one_scores

logical indicating if rules scores for each class should be calculated as a mean of one-vs-one scores instead of one-vs-rest manner. Default is FALSE.

platform_wise_scores

logical indicating if rules scores for each class should be calculated in each platform-wise then averaged instead of merging all platforms together. Default is FALSE.

disjoint

is a logical value indicating whether only disjoint pairs should be considered in the final set of selected pairs; i.e. all features occur only once among the set of TSPs. This is an argument to be passed to the training function SWAP.Train.KTSP from switchBox package.

seed

an integer to set a seed for the training process (for reproducibility).

classes

optional vector contains the names of classes in the wanted order. This means the individual classifiers will be ordered based on this vector. If this vector does not have all class names, then no classifiers will be train for those classes that are not mentioned and the samples from these classes will be removed from the training dataset.

SB_arg

list of any additional arguments to be passed to the training function SWAP.Train.KTSP from switchBox package

verbose

a logical value indicating whether processing messages will be printed or not. Default is TRUE.

Details

This function uses SWAP.Train.KTSP function from switchBox where the algorithm (Afsari et al (AOAS, 2014)) chooses the optimal number of rules (i.e. pairs) among the input range.

Value

Returns OnevsrestScheme_TSP object which contains one-vs-rest classifiers for the classes. These individual classifiers are top score pairs classifiers.

Author(s)

Nour-al-dain Marzouka <nour-al-dain.marzouka at med.lu.se>

Examples

# random data
Data <- matrix(runif(10000), nrow=100, ncol=100,
               dimnames = list(paste0("G",1:100), paste0("S",1:100)))

# labels
L <- sample(x = c("A","B","C"), size = 100, replace = TRUE)

# study/platform
P <- sample(c("P1","P2"), size = 100, replace = TRUE)

object <- ReadData(Data = Data,
             Labels = L,
                   Platform = P)

# not to run
# switchBox package from Bioconductor is needed
# Visit their website or install switchBox package using:
# if(!requireNamespace("switchBox", quietly = TRUE)){
#       if (!requireNamespace('BiocManager', quietly = TRUE)) {
#       install.packages('BiocManager')
#      }
#      BiocManager::install('switchBox')", call. = FALSE)
#  }

#filtered_genes <- filter_genes_TSP(data_object = object,
#                                  filter = "one_vs_rest",
#                                  platform_wise = FALSE,
#                                  featureNo = 10,
#                                  UpDown = TRUE,
#                                  verbose = FALSE)

# training
# classifier <- train_one_vs_rest_TSP(data_object = object,
#                              filtered_genes = filtered_genes,
#                              k_range = 10:50,
#                              include_pivot = FALSE,
#                              one_vs_one_scores = FALSE,
#                              platform_wise_scores = FALSE,
#                              seed = 1234,
#                              verbose = FALSE)

# results <- predict_one_vs_rest_TSP(classifier = classifier,
#                                   Data = object,
#                                   tolerate_missed_genes = TRUE,
#                                   weighted_votes = TRUE,
#                                   verbose = FALSE)

# Confusion Matrix and Statistics on training data
#  caret::confusionMatrix(data = factor(results$max_score, levels = unique(L)),
#                         reference = factor(L, levels = unique(L)),
#                         mode="everything")

# plot_binary_TSP(Data = object, classes=c("A","B","C"),
#                 classifier = classifier,
#                 prediction = results,
#                 title = "Test")