Witt: Querying Technology Terms Based on Automated Classification

Abstract

Witt is a tool that systematically and automatically categorizes software technologies using original information extraction algorithms applied to Stack Overflow and Wikipedia. Witt takes as input a term, such as "django", and returns one or more categories that describe it (e.g., "framework"), along with attributes that further qualify it (e.g., "web-application"). Our comparative evaluation of Witt against six independent taxonomy tools showed that, when applied to software terms, Witt has better coverage than alternative solutions, without a corresponding degradation in the number of spurious results. The information extracted by Witt is available through the Witt Web Application, which allows users to query and explore Witt's categorization of software technologies by both obtaining the category for a term, and all the terms in a given category. On-line Portal: https://cs.mcgill.ca/~swevo/witt-web Video: https://www.youtube.com/watch?v=tPsp1M4Ua3w