gen_fdr_power_curve: Generate precision-recall curve

In liangyy/silver-standard-performance: Generating PR/ROC curves in light and flexible way

Description

Given the set of true signals and the set of candidate genes along with their scores and proposing method, generate precision-recall curve, which is a sequence of precision and recall measured at a set of cutoffs

Usage

gen_fdr_power_curve(true_genes, gene, score, method = "gt", cutoff = NULL)

Arguments

true_genes	the set of true signals
gene	the set of all candidate signals
score	the scores of the candidates (the same length as gene)
method	method = 'gt' to propose score greater than threshold; method = 'lt' to propose score less than threshold
cutoff	cutoff = NULL will use all unique scores from true signals as cutoffs (essentially this is the finest grid we can get)

Value

a data.frame with all columns of compute_power_and_fdr's output along with cutoff column telling which cutoff is used to obtain the results of the row and precision (same as 1 - fdr) and recall (same as power)

Examples

gen_fdr_power_curve(
  true_genes = c('g1', 'g3', 'g4'),
  gene = c('g1', 'g2', 'g3', 'g4', 'g5', 'g6'),
  score = 1:6,
  method = 'gt',
  cutoff = NULL
)

liangyy/silver-standard-performance documentation built on Dec. 5, 2019, 8:53 a.m.