A visualization method for text data by Self-Organizing Maps

- Analysis of news texts which are related to the Great East Japan Earthquake-

Abstract

In this paper, we propose a novel visualization method for large amount of text data which are related to a big disaster such as the Great East Japan Earthquake. Firstly, we were classified articles on the basis of occurrences of words by Self-Organizing Maps (SOM). Then, we determined topics on the SOM from the Z-score of word occurances in each node. Six topics were extracted from 14019 articles of the Great East Japan Earthquake.[This abstract is not included in the PDF]