Extracting Relations among Search Properties Based on the Operational Context of Geographical Information Retrieval Systems

摘要

During searches for suitable objects with geographical information retrieval systems, users can select various search properties such as the region, category, price, distance, and rating. We assumed that there would be implicit connections among the properties in the search operations adopted by users. For example, if a user wants to go a bar, they might search for a bar by changing the rating property to "high" and changing the price property to "moderate." If the user wants to go a bar again, however, they might search for a bar by changing the rating property to "moderate" and changing the price property to "low." Thus, the price property and rating property are connected by the context when searching for the "bar" category. We consider that the connected properties may differ according to the operational context. In this paper, we propose a method for extracting connected properties based on the operation history of user. We also describe an automatic adjustment function based on the connected properties extracted.