Better Metrics to Predict the Performance of Personalized Web Search

Abstract

Our runtime generalization aims at striking a balance between two predictive metrics that evaluate the utility of personalization and the privacy risk of exposing the generalized profile. We present two greedy algorithms, namely Greedy DP and Greedy IL, for runtime generalization. We also provide an online prediction mechanism for deciding whether personalizing a query is beneficial. Extensive experiments demonstrate the effectiveness of our framework. The experimental results also reveal that Greedy IL significantly outperforms Greedy DP in terms of efficiency..