Analysis of Appropriate Category Level of Web Directory for Cross-Language Information Retrieval

In this paper, we analyzed appropriate category level of Web directory for Cross-Language Information Retrieval (CLIR). Our proposed method for CLIR is based on estimating domains of the query using hierarchic structures of Web directories. Therefore, it is necessary for correct domain estimation to detect appropriate category level of Web directory. We conducted experiments of retrieval using four category level in order to detect appropriate category levels of Web directory. We found that 2nd lv or 3rd lv is appropriate for CLIR.