Table based AHC algorithm for clustering words
Taeho Jo · 2016 18th International Conference on Advanced Communication Technology (ICACT) · 2016
This research proposes the table based AHC algorithm as the approach to the word clustering task. The results from encoding texts into tables were successful in the previous works on the text categorization and the text clustering, and if oppositely to the case of the text encoding, texts are assumed to be elements of each word, it becomes to be possible to encode words into tables. In this research, we map words into tables in the reverse direction to the text encoding, modify the AHC algorithm into the version where tables are given as its input, and apply the modified version of AHC to the word clustering task. In this research, we expect the better and more stable clustering performance, than the traditional version. Hence, this research provides the improved method of clustering words semantically.