A collection of implementations of popular machine learning algorithms using complicated data structures / metrics / methods and other improvements.
This is primarily a training project, in which experiments will be carried out to improve the efficiency of existing algorithms and implement their various variations. Successful examples can be useful for junior data scientists in their work and research.
All algorithms are implemented in Python, using numpy, scipy and autograd.
- [K-nearest neighbors] (Simple kNN algorithm, but using a k-dimensional tree data structure inside to speed up nearest neighbor searches)