evaluate_partition_semisup: Evaluation of a semisupervised obtained partition in...

In LilithF/doMIsaul: Do Multiple Imputation-Based Semi-Supervised and Unsupervised Learning

Description

Evaluate number of clusters, ARI, AUC difference, c-index and CPE, with Supervised, unsupervised and Semisupervised reference partitions

Usage

evaluate_partition_semisup(
  partition,
  ref.unsup,
  ref.sup,
  ref.semisup,
  data.surv,
  TMIN = 2,
  TMAX = 5
)

Arguments

partition	Vector containing cluster ids of the partition to evaluate.
ref.unsup	Vector: Unsupervised reference partition (i.e. data structure).
ref.sup	Vector: Supervised reference partition (i.e. using survival parameters).
ref.semisup	Vector: Semisupervised reference partition (i.e. combining both).
data.surv	dataframe with variables time and status.
TMIN	time point to start analyzing AUC.
TMAX	time point to analyze AUC.

Value

a list of named performances values

Examples

library(survival) # survival should be loaded in the environment
data(cancer, package = "survival")
cancer$status <- cancer$status - 1
if (requireNamespace("CPE", quietly = TRUE)) {
  res <- evaluate_partition_semisup (
    partition = factor(rep(c(1,2,3), each = 50)),
    ref.unsup = factor(rep(c(1,2,3), times = c(100, 25, 25))),
    ref.sup = factor(rep(c(1,2), times = c(50, 100))),
    ref.semisup = factor(rep(c(3, 2, 1), times = c(120, 10, 20))),
    data.surv = cancer[1:150, c("time", "status")] )
} else {
  print("Partition evaluation skipped as CPE is not available")
}

LilithF/doMIsaul documentation built on Dec. 17, 2021, 12:08 a.m.