Personalizing Search Based on user Search Histories (original) (raw)

In improving the quality of various search services on the Internet, Individualized web search (IWS) has demonstrated its effectiveness. User preferences are modelled as hierarchical user profiles in IWS applications. We propose a IWS framework called UPS that can adaptively generalize profiles by queries. Our runtime generalization evaluates the utility of personalization and the privacy risk of exposing the generalized profile. We present two greedy algorithms, namely GreedyDP and GreedyIL, for runtime generalization. For deciding whether personalizing a query is beneficial, we also provide an online prediction mechanism. The experimental results also reveal that GreedyIL significantly outperforms GreedyDP in terms of efficiency.