Context-Relevant Term Search

Summary

This was a research project done during the course of CS4501, Information Retrieval in Fall of 2017 with Eric McCord Snook. The purpose of this research was to develop a system for retrieving context-relavant definitions of a term given the context.

Markdown Image

Example use case of the context-relevant term search

Research

The developed system was able to perform significnatly, in that the proper definition was listed within the first two in average. The Jelinek-Mercer smoothed language model was used as the main system, and others like Apache OpenNPL library’s Part of Speech tagging to increase the accuracy of the system. The detailed research paper can be found on the bottom of the page.

Markdown Image

Graph showing performance of the system

Markdown Image

Table showing performance of the system

[Research Paper](/docs/Context-Relevant Term Search.pdf)

The Downloadable pdf contains more detail on the research