A Classification of a Scene in a Field note Using Topic Model

Abstract

In this study, we propose a method that can represent and analyze the spatial features of scenes in historical materials (field notes) related to area studies. To promote area studies research, we introduce a method to construct text database of area research resources using semantic web technologies. To improve accessibility and deepen the understanding of an area using a field note, we also introduce Latent Dirichlet Allocation (LDA) method. We constructed a text database using a field note written by Yoshikazu Takaya, a prominent researcher in Southeast Asian area studies. We show an experimental result on detected 30 topics from the constructed database. In this paper, we inspect the detection results and describe the advantages of the proposed method.