perturb_dataset: Perturbate a boolean matrix

In redsnic/CIMICE: CIMICE-R: (Markov) Chain Method to Inferr Cancer Evolution

Perturbate a boolean matrix

Description

Given a boolean matrix, randomly add False Positives (FP), False Negatives (FN) and Missing data following user defined rates. In the final matrix, missing data is represented by the value 3.

Usage

perturb_dataset(dataset, FP_rate = 0, FN_rate = 0, MIS_rate = 0)

Arguments

dataset	a matrix/sparse matrix
FP_rate	False Positive rate
FN_rate	False Negative rate
MIS_rate	Missing Data rate

Details

Note that CIMICE does not support dataset with missing data natively, so using MIS_rate != 0 will then require some pre-processing.

Value

the new, perturbed, matrix

Examples

require(dplyr)

example_dataset() %>%
  make_generator_stub() %>% 
  set_generator_edges(
    list(
      "D", "A, D", 1 , 
      "A", "A, D", 1 , 
      "A, D", "A, C, D", 1 , 
      "A, D", "A, B, D", 1 , 
      "Clonal", "D", 1 , 
      "Clonal", "A", 1 , 
      "D", "D", 1 , 
      "A", "A", 1 , 
      "A, D", "A, D", 1 , 
      "A, C, D", "A, C, D", 1 , 
      "A, B, D", "A, B, D", 1 , 
      "Clonal", "Clonal", 1 
  )) %>% 
  finalize_generator %>% 
  simulate_generator(3, 10) %>% 
  perturb_dataset(FP_rate = 0.01, FN_rate = 0.1, MIS_rate = 0.12)

redsnic/CIMICE documentation built on March 30, 2022, 2:46 a.m.