EasyHardClassifier-class: Container for a Pair of Trained Classifiers

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

Description

Stores two classifiers and the ID of the data set each one of them was trained on. Not intended for end-user.

Constructor

EasyHardClassifier(easyClassifier, hardClassifier, datasetIDs)

Creates a EasyHardClassifier object which stores the details of the two underlying classification models.

easyClassifier

A classifier trained on the easy-to-collect data set by easyHardClassifierTrain. Represented as a list of rules, each of which is also a list.

hardClassifier

A list of two with names "selected" and "model". The "selected" element should contain the chosen features and the "model" element is for the trained model on the hard-to-collect data set or simply a character vector of length 1 containing a class name, if the prediction of the samples left over from the easy classifier all or all except one belonged to a particular class.

datasetIDs

A vector of length 2 with names "easy" and "hard" containing the data set IDs from the original MultiAssayExperiment input data object for the easy data set and the hard data set.

Author(s)

Dario Strbenac

Examples

  predictiveRules <- list(list(feature = "age", relation = "<", value = 18, predict = "Risk"))
  hardClassifier <- DLDAtrainInterface(matrix(rnorm(400), ncol = 20),
                        classes = factor(sample(c("Safe", "Risk"), 20, replace = TRUE)))
  EasyHardClassifier(predictiveRules, list(selected = LETTERS[1:5], model = hardClassifier),
                     setNames(c("clinical", "RNA-seq"), c("easy", "hard")))

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