Utilizing Web Directories for Translation Disambiguation in Cross-Language Information Retrieval

In this paper, we analyzed the appropriate category level for a Web directory for cross-language information retrieval (CLIR). Our proposed method for CLIR is based on an estimation of the domains of a query using the hierarchical structures of Web directories. Therefore, correct domain estimation is necessary for detecting the appropriate category level of the Web directory. We conducted retrieval experiments using four-category levels to find the appropriate category levels of a Web directory. We found that the 2nd or 3rd levels the most appropriate for CLIR.