Divide training , validation and testing data.
52 views (last 30 days)
Show older comments
Sujith Jacob on 27 Jun 2022
Answered: Image Analyst on 27 Jun 2022
How can I divide only training and validation data randomly and have a separate contingous block for testing data.
for eg. if I have 2000 target points. I want to have randomly selected points from first 1500 points for training and validation but for testing I want 1501 to 2000 target points.
A = rand(2000,3) ; % your data
Test = A(1501:end,:) ; % take test continuously
A = A(1:1500,:) ; % pick the left data
A = A(randperm(1500,1500),:) ; % randomise the data
train_idx = round(70/100*1500) ; % 70% training
Train = A(1:train_idx,:) ;
Valid = A(train_idx+1:end,:) ;
More Answers (1)
Image Analyst on 27 Jun 2022
Depends on what kind of network training you're doing. If you're using trainNetwork and labels, then you can use imageDatastores and the function splitEachLabel
% Split the image data store into 80% for training, 10% for validation, and 10% for testing.
[trainingSet, validationSet, testSet] = splitEachLabel(imds, 0.8, 0.1);
Find more on Sequence and Numeric Feature Data Workflows in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!Start Hunting!