Document retrieval based on word co-occurrences extracted from educational information resources on the Internet

摘要

Currently it is difficult to easily find usable educational contents from a lot of various information resources on the Internet. In this article, we propose a new method to expand the user's query, based on word co-occurrences extracted from classified educational documents. In our recent experiment using newspaper articles, a single-word query with several additional major co-occurrent words showed better result with higher precision, in a case of 7 subcategories for high school science domain. This tendency seems prominent in lower recall part of the results, and will contribute to efficient educational information retrieval.