how to generate indices for the training data for all the possible combinations in leave n out cross validation?

I am planning on doing a leave n out cross validation. I would like to extract indices from the given data for all the possible combinations of leave n out process. For ex:- if I have 49 data points. and I do leave one out. Then the final outcome should have 49!/(48!*1!)=49 combinations of indices in this process. Is there a matlab function which would help with this issue?

Answer 1

[Athul Prakash]

Hi Harshan,

You may use 'cvparition' function with 'leaveout' option specified.

Have a read of the doc:

https://www.mathworks.com/help/stats/cvpartition.html#d122e224716

Hope it helps!