Method for supporting analysis of personal relationships through place names extracted from documents

Visualizing information extracted from text is helpful for intuitively understanding the information. Extracting and visualizing personal relationships from text is one of the promising applications of this approach. Existing methods usually estimate personal relationships from direct co-occurrences of personal names that appear in a text. In our previous work, we proposed a method for extracting personal relationships from indirect co-occurrence relationships obtained through place names. This method can estimate the relationships among persons who do not necessarily have direct relationships. These relationships are visualized in a network graph. However, it becomes difficult to grasp the relationships when the number of persons increases. In this paper, we propose a method that supports analyzing the extracted personal relationships through place names and that is based on our previous work. Our goal is to support analysis by providing the information of the clustering of closely related people and important place names for each cluster. The proposed method was applied to a Japanese historical chronicle written in the 12th century. Experimental results showed a strong correspondence to the known historical facts. The results also indicate that the proposed method might be able to uncover the characteristics of people whose histories are not clearly known yet.