Feature Selection using Swarm Optimization

COMP 1252 MSc Project
Associated with University of Greenwich

Contributors

Syed Md. Irfanul Alam, Mohammed Majid Al-Rifaie

Description
  • The problem with high dimensional data is that it has unnecessary features for classification
  • The aim of the project is to reduce dimensionality by removing unncessary features and improve classification accuracy
  • Implemented principal componant analysis and swarm intelligence techiniques including that of Particle Swarm Optimization and Dispersive Flies Optimization for feature selection
  • Reduced dimensionality from 2400 features to 677 features and improved classification accuracy by 2%

Technologies

Python, Jupyter, Scikit-Learn

Github Repo

swarm-intelligence

Source Document

Not Published