A machine learning engineering system from Sofia university students.

The problem

The problem we are trying to solve with this project is the slowness of weka. The project manages to make some of the algorithms not only faster but also with better performance(accuracy). The system also tries to be highly extensible, so you can plug in your own algorithms easily.

Currently we have implemented the following:

Clustering algorithms

• ZeroR
• K-Means

Classification algorithms

• K Nearest Neghbor
• Two Layer Perceptron
• ZeroR
• OneR
• Decision Tree

Classifiers result analyzer

• Cross-Validation
• Wilcoxon's Test
• Paired T Test

Data Preprocessors

• NormalizeValidator

Documentation

Every one of these components has extensive documentation and we really try hard to keep it this way!