GSoC/GCI Archive
Google Summer of Code 2010 Ushahidi

A predictive tagger module to tag short text reports using NLP & Active Learning techniques.

by Nishith Subiet Rastogi for Ushahidi

This module aims to tag the short text reports (SMS, Tweets, E-mail subject lines) aggregated by SwiftRiver or Ushahidi with suitable keywords using Natural Language Processing with Active Learning techniques. The predictive tagger will be built upon a proper-noun centric approach and utilizes the association between various keywords based on past reports to generate suitable tags for future reports.