INFORMATION RETRIEVAL USING COMMUNITY ONTOLOGIE

Abstract

It is unquestionable that "risk communication" is required for consensus building on many risks of technologies. The communication, however, can be obstructed by unfamiliarity and difficulty of technical terms or technical concepts which experts have developed for a long time. To improve accessibility to the technical information, we propose a novel technique for information retrieval (IR) which uses community ontology of experts. We will detail the notion of community ontology for formulating ontology vector, an ontology-oriented feature vector to introduce it into the conventional vector IR model. Then we will show results of an experiment demonstrating improvement of performance brought by the proposed technique, and discuss the effect that will be produced by the technique in practical use.