qtlscr.rank: Interactions by 95th percentile of normalized high order...

In yifansha/highinthunt: High Order Interaction Variable Importance Through Random Forest

Description

This function output the interaction terms of HIvimp object by 95th percentile of normalized high order interaction variable importance density estimation.

Usage

qtlscr.rank(o,data.nm,
            qtlcut=0.95,plot=TRUE)

Arguments

o	An object of class HIvimp.
data.nm	Data name provided by user for output and plot presentation.
qtlcut	Quantile score cut to decide final interactions to be presented.
plot	Option to generate quantile score plots for output interactions shown in groups, where all other group elements are strict subset of each group identifier in red color. If this option is turned off, only output interactions with quantile scores will be generated.

Value

A list of interactions with their quantile scores. Only interactions with quantile scores less than (1-qtlcut) will be shown. Plots of output interactions in groups with group elements being strict subset of each group identifier in red color.

Author(s)

Yifan Sha and Min Lu

Examples

data(express)
formula <- y~.
o <- HIvimp(y~., data = data, method = "hivimp", terms = NULL, PMDorder=TRUE, 
             importance = "permute", block = 1,
             subRF = T, subvars = NULL,
             w.initial = "vimp",
             wt = function(btpmd,digpmd){log(1/digpmd/btpmd)},
             itrSub = NULL, wtSub = TRUE,
             itrWt = 1, subQtl = NULL, max.var = 50,
             nHIvimp = function(all.cmbn){min(length(all.cmbn),150)},
             qtlscr = TRUE,
             verbose = TRUE)
qtlscr.rank(o,data.nm="express")

yifansha/highinthunt documentation built on July 2, 2020, 6:29 p.m.