Welcome to PyRidge’s documentation!

This project is aimed to write an useful machine learning library in Python based on the Ridge classification algorithms. Public repository is available in Github. These algorithms can be known in the literature also as Extreme Learning Machine, and they are explained as a type of a feedforward neural network where some neurons does not require to be tuned by calculating them. Algorithms are the same and they achieve to a reasonably good solution.

These supervised machine learning algorithms can be known in the literature as Ridge Classification, Tikhonov regularization or Extreme Learning Machine. A nice discussion about first and second terms can be seen in this discussion in StackExchange.

Although ELM is a polemic topic due to the accusations of plagiarism (see more here and here), some actual research is done by applying ensemble techniques to Ridge Classification, thus some some papers are used for implementing algorithms.

Main motivation of this repository is translating from MATLAB to Python 3 what I am doing in my PhD in Data Science in Universidad Loyola Andalucía.

The library is organized in the following way:

• Generic classes
  • Predictor
  • Neural Method
  • Kernel Method
• Predictors
  • Single predictors
  • Neural Network Ensembles
  • Negative Correlation Ensembles
• Functions for experiments and utilities
  • Utilities
  • Experiments