[1N3-J-9] Natural language processing, information retrieval: understanding

[1N3-J-9-01] Extract Object of Changes from Documents using Similarities of Co-occurrence Word and its Time Distribution

〇Katsuaki TANAKA¹ (1. Saitama Institute of Technology)

Keywords:document stream, text mining

In this paper, we proposed document sets consists of two types, a diversity description type that records actions to different objects and a change description type that recordings actions to objects that could be regarded as the same and attempted to extract objects of a change description documents. We assumed that co-occurrence words of a word indicating the object would have high similarities and words having different appearance time distributions in change description documents. As a result, we confirmed that it is possible to extract objects of a change description type documents.