Profile Based Personalized Web Search

Decent Essays
We suggest a Profile-based personalized web search framework UPS (User customizable Privacy-preserving Search), for each query ac-cording to user specified privacy requirements profile is generated. For hierarchical user pro-file we trust two conflicting metrics, namely personalization utility and privacy risk, with its NP-hardness proved we formulate the problem of Profile-based personalized search as Risk Profile Generalization.

With the help of two greedy algorithms, namely GreedyIL and GreedyDP, we generate the expected search result, greedy algorithms support runtime profiling. While the former tries to maximize the discriminating power (DP), the latter attempts to minimize the in-formation loss (IL). By exploiting a number of heuristics, GreedyIL outperforms GreedyDP significantly.

For the client to decide whether to personalize a query in UPS we provide an inexpensive mechanism. Before each runtime profiling this decision can be made to improve the stability of the search results while avoid the needless exposure of the profile.

3.1 System Architecture :

Indeed, the privacy concern is one of the main barriers is how to attain personalized search though preserving users privacy and deploying serious personalized search applica-tions. Hence we propose a client side profile-based personalization which deals with the preserving privacy and envision possible fu-ture strategies to fully protect user privacy. For

Fig. 1. Personalized Search Engine
Get Access