toyclassifiers

Motivation

Provide simple, unoptimized implementations of common classification algorithms to foster comprehensive understanding of the algorithms including practical (implementation) details and optimization possiblities.

Requirements

• Python 2.7
• Installation of scikit learn for running test script

Implementation Status

• Naive Bayes

  • only supporting continuous input variables

• Decision Tree

  • uses only entropy as scoring function

• K-Means

  • uses euclidean squared distance
  • uses random initalization or provided coordinates for centroids