error: [inputValues, targetValues] = input_preprocess();
Show older comments
[inputValues, targetValues] = input_preprocess(); % Transform the labels to correct target values.
% Choose form of MLP:
numberOfHiddenUnits = 1250;
% Choose appropriate parameters.
learningRate = 0.1;
% Choose activation function.
activationFunction = @logisticSigmoid;
dActivationFunction = @dLogisticSigmoid;
% Choose batch size and epochs. Remember there are 60k input values.
batchSize = 20;
epochs = 10000;
fprintf('Train twolayer perceptron with %d hidden units.\n', numberOfHiddenUnits);
fprintf('Learning rate: %d.\n', learningRate);
[hiddenWeights, outputWeights, error] = train(activationFunction, dActivationFunction, numberOfHiddenUnits, inputValues, targetValues, epochs, batchSize, learningRate);
save Train.mat hiddenWeights outputWeights
1 Comment
faaruuq jamaludin
on 12 Mar 2018
Answers (0)
Categories
Find more on Pattern Recognition and Classification 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!