RATE: Predictor Prioritization via RelATive cEntrality (RATE)...

In lcrawlab/SINATRA: Sub Image Analysis using Topological Summary Statistics (SINATRA)

Predictor Prioritization via RelATive cEntrality (RATE) centrality measures ###

Description

NOTE: This function assumes that one has already obtained (posterior) draws/estimates of a nonparametric or nonlinear function as suggested in Crawford et al. (2018) ###

Usage

RATE(
  X,
  f.draws = NULL,
  pre.specify = FALSE,
  beta.draws = NULL,
  prop.var = 1,
  rank.r = min(nrow(X), ncol(X)),
  nullify = NULL,
  snp.nms = NULL,
  cores = 1
)

Arguments

X'	is the nxp design matrix (e.g. genotypes) where n is the number of samples and p is the number of dimensions. This is the original input data
f.draws'	is the Bxn matrix of the nonparametric model estimates (i.e. f.hat) with B being the number of sampled (posterior) draws;
nullify'	is an optional vector specifying a given predictor variable effect that user wishes to "remove" from the data set. An example of this corresponds to Figures 2(b)-(d) in Crawford et al. (2018);
snp.nms'	is an optional vector specifying the names of the predictor variables;
cores'	is a parameter detailing the number of cores to parallelize over. If too many are assigned, RATE will set this variable to maximum number of cores that are available on the operating machine.

lcrawlab/SINATRA documentation built on Sept. 13, 2023, 2 p.m.