Abstract
Powerful methods for exploration and search from collections of documents are needed to manage the ever-increasing flood of digital information. In this paper, we propose an application of Self-Organizing Maps (SOM) to Information Retrieval. At first, words in "KOUJIEN" are encoded based on their meaning, and then create Word SOM. Documents are encoded using Word SOM, and then create Documents SOM. As a result of that, similar documents lie near each other on the map. We expect that not only the most required document but also relative documents can be found.
