Searcher's Activity in Standalone and Web Applications as a Source for Search Query Expansion

Abstract

The Web has so much variable information, therefore searching a specific one is complex task. To find something valuable, specifying good search query is crucial. However, average search query consists of only about two words which cannot specify intent of a searcher well. We suppose that these words follow in most cases the searcher's activity. Based on this assumption we propose an approach for search query extension by the searcher activity context. The activity context contains words reflecting the searcher's activity context expressed by keywords gathered from the content of one of the very recent activities within applications provided by the searcher. We present an approach for finding out searcher's activity context by analyzing the content and his interaction between applications considering both standalone applications and applications running inside a browser. Our method uses the activity logs to find a connection between the search query and specific application. Then we extend the query by terms gathered by analyzing the selected application content. We evaluate our approach in a series of experiments based on data gathered by monitoring a group of searchers by means of developed logger prototype.