Hill Climbing for Diversity Retrieval

Case-based recommender systems have been widely applied in suggesting products that are most similar to current user´s query. By prioritizing similarity during a case-based approach may degrade the quality of the retrieval results. There have been a number of attempts to increase retrieval diversity. However, there is a trade-off between similarity and diversity. The improvements in diversity may lead to the loss of similarity. In this paper, we propose a new retrieval strategy based on the random-restart hill-climbing algorithm which optimizes the trade-off between similarity and diversity. Experimental results show that the proposed algorithm can achieve a better overall quality than other approaches.