Create an index ind with the column numbers of the predictors that were “statistically significantly” associated with y. Use a p-value cutoff of 0.01 to define “statistically significant”. How many...

Create an index ind with the column numbers of the predictors that were “statistically significantly” associated with y. Use a p-value cutoff of 0.01 to define “statistically significant”. How many predictors survive this cutoff?

Re-run the cross-validation but after redefining x_subset to be the subset of x defined by the columns showing “statistically significant” association with y. What is the accuracy now?