regularized_linear_regression: Compute regularized parameter estimates for the linear model

In lkshrsch/gLVInterNetworks: Inference of interaction networks based on generalised Lotka Volterra dynamics

Usage

regularized_linear_regression(X,Y, alpha, cvfolds)

Arguments

X

Examples

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.

## The function is currently defined as
function(X,Y, alpha, cvfolds){
  
  
  #require(glmnet)
  
  sol = matrix(0, nrow = nrow(Y), ncol = nrow(X))
  
  # X <- X[-1,]  # Get rid of intercept , it is automatically placed by glmnet
  
  
  mse = 0
  
  
  
  for(i in 1:nrow(Y)){
    
    
    cv <- glmnet::cv.glmnet(t(X)[,-1],Y[i,], alpha=alpha ,nfolds = cvfolds,standardize=T)
    
    small.lambda.index <- which(cv$lambda == cv$lambda.1se)
    small.lambda.betas <- cv$glmnet.fit$beta[,small.lambda.index]
    small.lambda.betas <- c(cv$glmnet.fit$a0[small.lambda.index],small.lambda.betas)
    sol[i,] = small.lambda.betas
    mse[i] = cv$cvm[small.lambda.index]
    #print(cv$lambda.1se)
  }
  
  mse = mean(mse)

  return(list(sol,mse))
  
}

lkshrsch/gLVInterNetworks documentation built on May 21, 2019, 7:33 a.m.