Question

We are required to find the best value of k for k-NN classification on a data set of size 20,000. When checking for different values of k using 5-Fold cross validation we divide the training set into 5 partitions. (a). What is the size of each of the training set used for 5-Fold cross validation? (b). What is the estimated error for 4-NN classification using 5-Fold cross validation if the summary of errors is as under:

Answer 1

In 5-Fold cross validation, each training set would have a size of 4,000. The estimated error for 4-NN classification using 5-Fold cross validation can be calculated with the provided summary of errors.

Step-by-step explanation:

(a). To find the size of each training set used for 5-Fold cross validation, divide the total size of the data set (20,000) by the number of folds (5). Each training set would have a size of 4,000.

(b). To estimate the error for 4-NN classification using 5-Fold cross validation, the summary of errors is needed.