Twitter Noun-Verb Relation Thesaurus Tree

Naoki Urabe, Jun Rokui · 2018

The thesaurus is a representation of words and concepts in tree structure[1]. Methods for automatically constructing a thesaurus are being actively studied. Thesauri constructed using conventional methods cannot deal with changes in word usage. Therefore, in this research, we build a thesaurus applying the C4.5 algorithm for Twitter data. The word use varies depending on when the Tweet data was generated. Therefore, we can capture the changes in how words are used. By applying the C4.5 algorithm to Twitter data, it is possible to efficiently derive the relationship of words. We verified the effectiveness of the Twitter thesaurus.

Read the paper · More papers on PaperTik