How do I partition data sets for cross validation in MATLAB?

I'm trying to split, or partition, the data into two groups. Testing Data and Training Data. Ideally I want to write a function that can randomly divide the data into a variable sized patition. So that I could do specifi and leave one out cross validation. I'm not sure how I'll do this though.

Answer 1

[Swetha Polemoni]

Hi

It is my understanding that you want to partition the dats randomly. You can use either cvpartition from Statistics and Machine Learning Toolbox or randperm to serve the purpose.

The following answer might help you to understand how to use 'randperm'

how to divide a data set randomly into training and testing data set?

Answer 2

[Enrico]

I am trying to use the cvpartition object but I'm struggling understanding its main structure.

Based on what I know, when I want to perform a CV partition, let's say with 10 folds, I'm supposed to make these folds randomized. however when I create a cvpartition object (e.g. cvpartition(200000, "KFold",10), I get a result that looks like:

K-fold cross validation partition

NumObservations: 200000

NumTestSets: 10

TrainSize: 180000 180000 180000 180000 180000 180000 180000 180000 180000 180000

TestSize: 20000 20000 20000 20000 20000 20000 20000 20000 20000 20000

IsCustom: 0

However, I can't understand how are these training and test sets organized. I mean I was expecting have some place that tells me what indexes are in each training and test folds and yet I can't see where they are.

If I create two cvpartitions, with the same data and try to compare them I obtain:

isequal(cv1, cv2)

ans =

logical

0

However, I can't find their differences. The structure that defines them seems the same.