Further Experiments in Case-Based Collaborative Web Search

Collaborative Web Search (CWS) proposes a case-based approach to personalizing search results for the needs of a community of like-minded searchers. The search activities of users are captured as a case base of search cases, each corresponding to community search behaviour (the results selected) for a given query. When responding to a new query, CWS selects a set of similar cases and promotes their selected results within the final result-list. In this paper we describe how this case-based view can be broadened to accommodate suggestions from multiple case bases, reflecting the expertise and preferences of complementary search communities. In this way it is possible to supplement the recommendations of the host community with complementary recommendations from related communities. We describe the results of a new live-user trial that speaks to the performance benefits that are available by using multiple case bases in this way compared to the use of a single case base.