.NetML

Librarie containing machine learning models written in C# (.NET Framework 4.6.1).

Getting started

1. Feedforward neural network - model able to learn with backpropagation algorithm. Example:

var neuralNetwork = new FeedForwardNeuralNetwork(numberOfInputs, numberOfOutputs, hiddenLayers);
neuralNetwork.SetWeights(randomNumbers);
neuralNetwork.BackPropagationTrain(traningData, maxEpochs, learningRate);
double accuracyAfterTraning = neuralNetwork.GetAccuracy(testData);
var result = neuralNetwork.ComputeOutput(input);

Dependencies

Install-Package FileDeserializer -Version 1.0.0

TODO

• Random generated initial weights and biases.
• Other activations functions in neurons classes.
• Momentum

Sources

https://visualstudiomagazine.com/Articles/2015/04/01/Back-Propagation-Using-C.aspx?Page=1