Welcome to MOAFS’s documentation!

MOAFS is a library for the Massive Online Analysis (MOA) framework. It is based on the MOAReduction extension and contains seven feature selection algorithms to be used as dimensionality reduction techniques in data streams classification problems, especially in the text-domain field. MOAFS uses an incremental version of Naïve Bayes as the base classifier.

Getting started:

• Getting started
  • Installation
  • Dependencies
  • Sample datasets
  • Sample outputs
  • Running GUI
  • How to test this library
• Examples (command line)
  • Parameters
  • Classification without feature selection (No method)
  • Information Gain
  • Symmetrical Uncertainty
  • Chi-Squared
  • Cramers V-Test
  • Gain Ratio
  • Extremal Feature Selection
  • Online Feature Selection
  • Changes in Reduction rate
  • Changes in Processing window
• Examples (GUI)
  • Classification without feature selection (No method)
  • Information Gain
  • Symmetrical Uncertainty
  • Chi-Squared
  • Cramers V-Test
  • Gain Ratio
  • Extremal Feature Selection
  • Online Feature Selection
• API Documentation
  • Naïve Bayes
  • Information-based algorithms
  • Attribute Evaluator
  • Extremal Feature Selection
  • Online Feature Selection