This library implements various machine learning methods. The purpose here is mostly educational (for my own and hopefully others') and I aim for clarity rather than efficiency. Note that this is a work in progress! So far the library contains implementations for
- Multilayer Perceptrons
- Restricted Boltzmann Machines and Deep Belief Networks (with support for GPUs through gnumpy)
- k-Means clustering
- Principal Components Analysis
- Metric learning
- Method of Schultz, M., & Joachims, T. (2003) that learns a diagonal distance metric matrix
- A variant of Schultz, M., & Joachims, T. (2003) that learns a low-rank distance metric matrix