自己組織化マップを用いた類似文書抽出システムの開発(一般講演C4,自己組織化マップ)

抄録

In this paper, we describe a system to extract similar documents from a lot of these, using the Self-Organizing Maps (SOM). There is a feature that SOM is mapped a connection whose interrelation over multidimensional is rendered by one dimension or two dimensions. By using this feature, we analyzed the digitized vector by the feature of documents, and we created the SOM map. As a result, we succeeded in placing the similar documents in the SOM map, and showing the similarity between a document and another one.