Secure Personalized Recommendation System for Mobile User - Information Security and Cryptology-ICISC 2010 - page 267

looks very similar to an authentic user, it is a difficult task to correctly identify
shilling attacks. These attacks can cause the degradation of user trust in the objectivity
and accuracy of the recommender system.
Our system aims at providing personalized recommendation for mobile users based
on user profiles for tourism domain. Capturing user profiles naturally involves the
processing of personal data such as location data, personal preferences (interests),
travel information and so on. However, the processing of user data requires security
measures to ensure the user's fundamental rights to privacy. Current recommender
systems have the privacy problem [1]. For the privacy concerns, the user's personal
data must be protected and proceed in a safe manner. Trust concept can take
advantage over recommender system. Local trust, reputation, demographic trust and
location aware reputation [13] can be used to construct a trust model.
In [8], we proposed RPCF (Rule-based Personalization with Collaborative
Filtering) algorithm which can address the scalability, sparsity and cold-start problem
of pure collaborative filtering method and can give the accurate and good quality
recommendation to the user. In [9], we proposed the architecture of multi-agent
tourism system (MATS) which provides the most relevant and updated information
according to the user's interest by using RPCF Algorithm. In [7], we extended MATS
for mobile user.
This paper is an extension of the encyclopedia article [7] that gives the secure
personalized recommendation based on multi-agent technology in tourism industry to
serve the mobile users. This paper pays great attention to security issue. The privacy
and trust management are not considered in this paper. The primary contribution of
this paper is to detect or prevent the shilling attacks by adopting significant weighting
and trust weighting that complements to the RPCF algorithm for giving the accurate
recommendation to the user.
The rest of the paper is organized as follows. Section 2 describes the theoretical
background of the system and Section 3 presents the security architecture of the
personalized recommendation system in detailed. Section 4 points out the
experimental results of the system. Section 5 concludes with a summary and suggests
directions for future works.
2 System Background
This section describes the system background related with recommender system,
personalization system and the attack types.
2.1 Recommender System
Recommender system can be defined as a specific type of information filtering (IF)
technique that attempts to present information items (movies, music, books, news,
images, web pages, etc.) that are likely of interest to the user. The goal of a
recommender system is to generate meaningful recommendations to a collection of
users for items or products that might interest them.
Suggestions for topics on Amazon , or movies on Netflix , hotel recommendation on
Tripadvisor are real-world examples of the operation of industry-strength