Table based KNN for extracting keywords

Abstract

This research is concerned with the table based KNN as the approach to the keyword extraction task. The keyword extraction task is viewed as an instance of word classification, and it is discovered that encoding words into tables improved the word categorization performance. In this research, words are encoded into tables and the correspondingly modified version of KNN is applied to the keyword extraction task. As the benefits from this research, like the case in the general word categorization, we expect the better performance in the keyword extraction, as the special word classification. Therefore, the goal of this research is to provide the better scheme of extracting keywords from each text.