K-nearest neighbour Classifier- Cross validation

Hi, I am working of classifiers. My problem is as follow:

train_labels = Nx1; train_data = NxD;

test_ labels = Mx1; test_data = MxD;

where, M>>N;

I am using Knn classifier and I have to find the classification accuracy by k-fold cross-validation. Any help?

[PIRC]

which is helpful for identify the optimal K value and avoid overfitting, underfitting issues

Answer 1

[Tom Lane]

This example illustrates k-fold cross validation for knn classification:

http://www.mathworks.com/help/stats/classification-using-nearest-neighbors.html

However, you ordinarily wouldn't divide your data into train/test sets if you were going to use cross-validation.