Abstract
While a Japanese full text search system has the 2-gram characters index as base in order to save the volumes of the index file, the additional higher-gram index is expected to improve the performance. This paper describes how to select the additional index terms from the contents of database. To select the terms, this paper defines the entropy of strings which expresses the ambiguity of strings. This paper also describes the specification, including string input and entropy hierarchy, of the prototype term selection system in detail. To compare and test the effectiveness of the entropy, we have established a prototype system. As a result, we got some benefits and future issues about the processed approach.
