Welcome to the documentation for our machine learning project!

Implementations

Neural Network Classifier

A neural network for use in classification problems.

• DenseLayer - A fully connected layer with user defined size and activation function

• Loss Functions - The loss function to use in the neural network

  • (currently only categorical cross entropy is supported).

• Activation Functions - The activation functions to use in the fully connected layers

  • (currently stable_softmax and leaky_relu is supported).

Decision Tree Classifier

A decision tree for use in classification problems.

• Node - A tree node used by the decision tree classifier, is either leaf or not.

• Impurtiy Criterion - The impurity function to use when decision whether to split nodes in the decision tree

  • (currently only gini_impurity and entropy are supported).