Ranking web pages by associating keywords with locations

Abstract

Many Web queries contain both textual keywords and location words. When answering such queries, the association between the textual keywords and locations in a Web page should be taken into account. In this paper, we present a new ranking algorithm for location-related Web search, which is called MapRank. Its main idea is to extract the associations between keywords and locations in Web pages and further use them to improve ranking effectiveness. We first determine map each keyword with specific locations and form a set of < keyword, location > pairs. Then, we compute the location-constrained score for each keyword and combine it into the ranking procedure. We conduct comparison experiments on a real dataset and use the metrics including MAP and NDCG to measure the performance of MapRank. The results show that MapRank is superior to previous methods with respect to different symbolic-location-related queries. © 2013 Springer-Verlag Berlin Heidelberg.