How to train a classifier and test it separately?

3 views (last 30 days)
Good day,
I have a beginner experience in using the Matlab, Neural Network Toolbox. My question is: I have some N x 22 feature matrix for training data. Where N is very large. In addition, there are three class labels (targets) corresponding to each of the N samples. I am wondering whether I can train a neural network (in general a classifier) using the training data only. So that I can use this trained classifier to test my test samples one-by-one during the testing phase.
To be precise, I want to save my trained classifier (neural network or SVM or k-NN). Later it should provide the predicted class label/target given a test sample (one-by-one in a for-loop). Also, I came across dividing the data (<http://www.mathworks.com/help/nnet/ug/divide-data-for-optimal-neural-network-training.html>) and similar community Q/A but I am still skeptical in order to use this for a single test sample.
Thank you for your time and any help is much appreciated.

Accepted Answer

Greg Heath
Greg Heath on 6 Jul 2014
You can control the data division via net.divideFcn and net.divideParam
help dividetrain
doc dividetrain
Hope this helps
Thank you for formally accepting my answer
Greg

More Answers (0)

Categories

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!