R: Recipes for Analysis, Visualization and Machine Learning

Performing leave-one-out-cross-validation to limit overfitting

We provide the framework of the code to perform leave-one-out-cross-validation for linear regression. You should be able to easily adapt this code for any other regression technique. The rationale and explanation presented under the previous recipe Performing k-fold cross-validation apply to this one as well.

How to do it...

To perform leave-one-out-cross-validation (LOOCV) to limit overfitting, follow the steps below:

1. Read the data:

   > bh <- read.csv("BostonHousing.csv")

2. Create the two functions shown as follows; we show line numbers for discussion:

   1 rdacb.loocv.reg <- function(df) { 2 mean.sqr.errs <- sapply(1:nrow(df), rdacb.loocv.reg.iter, df) 3 list("mean_sqr_errs"= mean.sqr.errs, "overall_mean_sqr_err" ...