A tool that provides you with machine learning algorithms, data preparation techniques and last but not least comparing different models on a metric you decide. The purpose of this project is double: from one site you can use the tools for your analysis, from the other you can learn how a particular technique works, being it all implemented from scratch, using just numpy, pandas and few other basic packages.
- SDAR, SDEM algorithms;
- pca (numerical, categorical with pandas is coming soon);
- covariance shift tool;
- kmeans (with kmeans++ init);
- data split (train, test, validation).
- finishing kmeans-
- testing kmeans and PCA on real world datasets