MATLAB: How to disable the random partition in cvpartition

Question

Hi,

I'm trying to build a classification model, where I'm using the cvpartition function to 'HoldOut' a certain percentage of train and test data.

In my design, I want to train 50% and tst the rest. But the problem I'm facing is, it splits the data randomly which I don't want it to happen. I have 9000 rows, I want the first 4500 for train and the next 4500 for test.

How do I solve this? Here's my code :

rng('default'); 
partition = cvpartition(total.status, 'HoldOut', 0.5);
traindata = total(training(partition), :);
testdata = total(test(partition), :);

Thank you.

Answer 1

Hi,

I think if you don't want randomized data you can partition the data yourself. Something like what i have written below.

traindata = total(1:4500, :);
testdata = total(4501:9000, :);