pheno_feat_sel: Feature selection

In borishejblum/phenotypr: Elastic-net logistic regression for phenotyping prediction using EHR data

Feature selection

Description

Elastic net logistic regression

Usage

pheno_feat_sel(surrogate, features, covariates = NULL, ...)

Arguments

surrogate	a vector of length n containing. Coerced to numeric.
features	a matrix of dimension n x m containing the features from which to select predictors
covariates	an optional matrix of dimension n x r containing additional potential predictors
...	further agruments not used here

Value

a list containing 2 components:

• logisticEN_model a glmnet object that can be used to predict external observation.

• sel_feata vector containing either the names of the selected features via cross valiation on the "class" measure type (through cv.glmnet) using the lambda.1se parameter.

borishejblum/phenotypr documentation built on May 2, 2022, 11:04 p.m.