Skip to content

Latest commit

 

History

History
21 lines (14 loc) · 1.05 KB

README.md

File metadata and controls

21 lines (14 loc) · 1.05 KB

Fbook_KeyphraseExtraction

Repository contains code for data collection and preliminary experiments for my research with Dr. Yudong Liu at WWU in the natural language processing lab.

Summary

Our research goal is to build NLP models that can identify complex word for non-native english speakers. We want to use these models to build language learning applications that are integrated in the user's browser to create contextual learning.

Originally, we pursued keyphrase extraction as a potential direction for our project. The data indicated that this is a good direction to pursue and so we've shifted towards complex word identification instead.

Key highlights

  • Scrapers that collect all the comments of a Facebook post and automatically scrolls down a user's feed
  • Scripts that organize the data and filters to relevant information
  • Implemented a keyphrase extraction model (TextRank) from scratch and experimented with changing the grammar used. For example, TextRank only uses noun phrases as grammar, I played around with adding verb phrases.