prediction.analysis: STEP 3 IN PNC ANALYSIS - ANALYSIS ON NETWORK Perform...

In EmilyLMorris/ImagingPred: Code for performing prediction of functional connectivity using brain imaging data

Description

STEP 3 IN PNC ANALYSIS - ANALYSIS ON NETWORK Perform prediction of connectivity using provided covariates.

Usage

prediction.analysis(
  networks,
  covariates,
  method = "SVM",
  missing.prop = 0.05,
  ID = NULL,
  newdata = NULL
)

Arguments

networks	list of files containing estimated adjacency matrix for each subject.
covariates	data.frame containing all covariates.
method	used to specify which approach to use for prediction; default option is "SVM" and"randomForest" is also available.
missing.prop	double value used to specify the threshold for filtering out covariates if they are missing more than this specified percentage.
ID	vector of subject IDs that corresponds to the covariate matrix, option to specify if not included in the covariate data.frame.
newdata	option to provide new data to perform the prediction.

Value

List of three items: predicted variables_selected; predicted is a vector of predicted values; variables_selected is a matrix where a 1 indicates variables (columns) included in the model for a give edge (rows); index_edges is the index of which edges were modeled, some edges may be removed due to lack of variability across subjects.

EmilyLMorris/ImagingPred documentation built on June 19, 2021, 9:33 p.m.