get.performance: Get performance of feature selection

In KnowGRRF: Knowledge-Based Guided Regularized Random Forest

Description

Comparing feature selected by algorithms and the ground truth in simulation

Usage

get.performance(set.truth, set.sel, set.all)

Arguments

set.truth	A vector of important features set in simulation (ground truth)
set.sel	A vector of important features selected by algorithms
set.all	A vector of all candidate features for feature selection

Value

returned a vector of feature selection performance including JI (ratio of intersect of two sets and union of two sets), TPR (percentage of correctly selected features in all true important features) and FPR (percentage of wrongly selected features in true non-important features)

Author(s)

Li Liu, Xin Guan

References

Guan, X., & Liu, L. (2018). Know-GRRF: Domain-Knowledge Informed Biomarker Discovery with Random Forests.

Examples

set.truth=1:10  ##true important feature from ground truth
set.sel=c(8:10, 95)  ##selected feature by an algorithm
set.all=1:100  ##all candidate features

get.performance(set.truth, set.sel, set.all)

KnowGRRF documentation built on May 2, 2019, 6:43 a.m.