Identifier: TFM:41
Authors: Benkaryouh, Youssef
Abstract:
A new P2P system based on synthetic users' profiles for privacy protection in web search engines while maintaining quality of service Youssef Benkaryouh Advised By Dr. J. Castellà-Roca y Dr. A. Viejo Dept. of Computer Engineering and Maths, Universitat Rovira i Virgili, Av. PaÏsos Catalans 26, E-43007 Tarragona, Catalonia, Spain Abstract Web Search Engines (eg Google, Yahoo, Bing, etc.) store informa- tion related to the queries made by users, this allows the creation of users' profiles to improve the quality of service provided (personalized results, suggestions, corrections, etc.). The information contained in the profile created by the search engine for the user, may uniquely identify her and relate her identity with sensitive and confidential quieries, and therefore her privacy right is compromised. Currently, there are several proposals in the literature that address the problem of users' privacy in WSE offering them alternatives that solve this problem. Schemes based on the collaboration of users (P2P or social networks), called multi-party, use queries generated by real users to obfuscate the real interests of the user, and thus solve the problem of detecting automatically generated queries used in the schemes single-party based on individual privacy protocol execution. Although multi-party schemes provide improvements over the single-party schemes at the type of queries used, which improves the level of privacy, a good quality of service in these schemes is not yet achieved. For this reason, a new P2P system is proposed to allow users to be grouped into different categories based on their profile allowing them to execute a privacy protection protocol while maintaining a good quality of service at the same time.
Repositori URV