6 Commits

Author	SHA1	Message	Date
Jonathan Baldie	c32bf3fe2b	Configure an Activation Function per hidden layer (#208) ... * ability to specify per-layer activation function * some tests for new addition to layer * appease style CI whitespace issue * more flexible addition of layers, and developer can pass Layer object in manually * new test for layer object in mlp constructor * documentation for added MLP functionality	2018-02-01 23:15:36 +01:00
David Monllaó	e83f7b95d5	Fix activation functions support (#163) ... - Backpropagation using the neuron activation functions derivative - instead of hardcoded sigmoid derivative - Added missing activation functions derivatives - Sigmoid forced for the output layer - Updated ThresholdedReLU default threshold to 0 (acts as a ReLU) - Unit tests for derivatives - Unit tests for classifiers using different activation functions - Added missing docs	2018-01-09 11:09:59 +01:00
David Monllaó	c4ad117d28	Ability to update learningRate in MLP (#160) ... * Allow people to update the learning rate * Test for learning rate setter	2017-12-05 21:09:06 +01:00
David Monllaó	b1d40bfa30	Change from theta to learning rate var name in NN (#159)	2017-11-20 23:39:50 +01:00
David Monllaó	de50490154	Neural networks partial training and persistency (#91) ... * Neural networks partial training and persistency * cs fixes * Add partialTrain to nn docs * Test for invalid partial training classes provided	2017-05-23 09:03:05 +02:00
David Monllaó	4af8449b1c	Neural networks improvements (#89) ... * MultilayerPerceptron interface changes - Signature closer to other algorithms - New predict method - Remove desired error - Move maxIterations to constructor * MLP tests for multiple hidden layers and multi-class * Update all MLP-related tests * coding style fixes * Backpropagation included in multilayer-perceptron	2017-05-18 00:07:14 +02:00