Document Proximity: Keyword Query Suggestion Based On User Location

Abstract

 The keyword suggestion in web search allows users to access relevant information without having to know how to express their queries accurately. Existing keyword suggestion techniques do not take into account user locations or query results. that is, the spatial proximity of a user to the results obtained is not taken into account in the recommendation. However, the relevance of search results in many applications (for example, location-based services) is known to correlate with their spatial proximity to the query sender. In this article, we designed a framework for keyword query suggestions that takes into account location. We propose a weighted keyword chart, which captures the semantic relevance between keyword queries and the spatial distance between the resulting documents and the user's location. The chart is scanned randomly, step by reset, to select keyword queries with the highest scores as suggestions. For our framework to be scalable, we propose a partitioning approach that goes beyond the basic algorithm up to an order of magnitude. The relevance of our framework and the performance of the algorithms are evaluated with real data.