multiPRC: Precision-recall curve for multi-class prediction

In davismcc/cardelino: Clone Identification from Single Cell Data

Precision-recall curve for multi-class prediction

Description

Precision-recall curve for multi-class prediction

Usage

multiPRC(
  prob_mat,
  simu_mat,
  marginal_mode = "best",
  cutoff = NULL,
  multiLabel.rm = TRUE,
  add_cut1 = FALSE
)

Arguments

prob_mat	Probability matrix for each cell to each component
simu_mat	The true identity of assignment from simulation
marginal_mode	A string for the mode to marginalize the column: best, second, or delta
cutoff	A list of cutoff; if NULL use all unique scores
multiLabel.rm	Logical value; if True, remove the samples with multiple labels
add_cut1	Logical value; if True, manually add a cutoff of 1

Value

A list with two components: df, a data.frame containing precision and recall values at various cutoffs and AUC, the overall AUC.

davismcc/cardelino documentation built on Nov. 19, 2022, 2:44 a.m.