F_beta: F_beta

In ebecht/hypergate: Machine Learning of Hyperrectangular Gating Strategies for High-Dimensional Cytometry

F_beta

Description

Compute a F_beta score comparing two boolean vectors

Usage

F_beta(pred, truth, beta = 1)

Arguments

pred	boolean vector of predicted values
truth	boolean vector of true values
beta	Weighting of yield as compared to precision. Increase beta so that the optimization favors yield, or decrease to favor purity.

Examples

data(Samusik_01_subset)
truth=c(rep(TRUE,40),rep(FALSE,60))
pred=rep(c(TRUE,FALSE),50)
table(pred,truth) ##40% purity, 50% yield
#' F_beta(pred=pred,truth=truth,beta=2) ##Closer to yield
F_beta(pred=pred,truth=truth,beta=1.5) ##Closer to yield
F_beta(pred=pred,truth=truth,beta=1) ##Harmonic mean
F_beta(pred=pred,truth=truth,beta=0.75) ##Closer to purity
F_beta(pred=pred,truth=truth,beta=0.5) ##Closer to purity

ebecht/hypergate documentation built on Feb. 4, 2024, 3:29 p.m.