To obtain separate predicted values for training and testing datasets in ANN models in MATLAB, you typically rely on built-in tools or apps that handle data division and prediction automatically.
If you're using the Neural Network Toolbox or the Neural Fitting App:
- The toolbox automatically splits your data into training, validation, and testing sets based on specified ratios.
- After training, it stores the indices of each subset in the tr structure (e.g., tr.trainInd, tr.testInd).
- You can retrieve predicted outputs for each subset using the trained network and these indices.
For example:
trainPred = net(x(:, tr.trainInd)); testPred = net(x(:, tr.testInd));
This logic is embedded in the app-generated scripts, so you don’t need to write it manually if you use the GUI.
