Host: The Japanese Society for Artificial Intelligence
Name : The 32nd Annual Conference of the Japanese Society for Artificial Intelligence, 2018
Number : 32
Location : [in Japanese]
Date : June 05, 2018 - June 08, 2018
Newspapers and other media often describe particular topics over some periods by giving a series of texts that introduces new information and corrects previous information. By analyzing such texts, we can understand how a topic changed over time and how the media treated it. However, analyzing such texts manually is time-consuming. To solve this problem, we propose a method for visualizing texts that describe a topic changing over time. For this purpose, we first generate an event information graph from such texts by using an exiting method. Then we transform it into two different graphs. The first graph shows changes in the topic that we obtain by separating the initial graph and comparing the resulting time-dependent graphs. The second graph depicts a summary of the topic that we obtain by extracting high-frequency words from the initial graph. The results of our experiment show that our method can visualize changes and important information described in the texts.
