Article Preview
TopIntroduction
Along with the rapid development of network technology, the amount of information on the Internet is expanding rapidly, which leads to the serious problem of information overload (Wu et al., 2021a; Lin et al, 2021), and in turn the bottleneck of end users’ effective use of information resources on the Internet (Wu et al., 2021b, 2021c, 2021d). Based on the specific information needs of users (such as points of interest, locations and preferences), personalized information retrieval can provide users with resources to meet their personalized needs, and then help users quickly obtain the target data from massive resources, thus it is an effective tool to solve the problem of information overload (Zhou et al., 2020; Zhang et al., 2020), and has attracted wide attention from both scientific and industrial communities. However, on the one hand, along with the rapid development of emerging network technologies such as cloud computing, the server of personalized information retrieval is becoming more and more untrusted, and has become the main source of end user privacy disclosure (Liu et al., 2018; Such & Natalia, 2018). On the other hand, in order to obtain accurate service results, personalized information retrieval requests issued by users to the server certainly would contain a large number of sensitive information (such as interests, preferences and locations). All of this information would be collected by the untrusted server, which is bound to pose a serious threat to the user privacy (Wu et al., 2021e; Liu et al., 2021; Saura et al., 2021). Therefore, the problem of privacy and security has become a major obstacle to the further development and application of personalized information retrieval services on the Internet (Wang et al., 2019; Hewitt & White 2021), and thus has become an important topic in the field of organizational and end user computing.
In this context, this paper focuses on the privacy protection of personalized information retrieval, and proposes an effective solution. To obtain personalized results, it is required for each user to not only report his current geographic location (by using query locations) and personal preference (by using preference profiles) to the untrusted server, but also report the content that he wants to obtain (by using query points of interest). Therefore, the user privacy that needs to be protected in personalized information retrieval mainly includes location privacy (which can be obtained by analyzing query locations), query privacy (which can be obtained by analyzing query interest points) and preference privacy (which can be obtained by analyzing preference profiles). To this end, this paper focuses on the comprehensive protection of all kinds of user privacy in personalized information retrieval, whose goal can be summarized as follows. According to the distribution features of all kinds of request data (including preferences, locations and interest points) related to personalized information retrieval, we aim to construct a basic framework for user privacy protection, so as to overcome the application limitations of existing technical methods in personalized information retrieval, i.e., to comprehensively improve the security of users’ preference privacy, location privacy and query privacy on the untrusted server, without compromising the availability of an existing personalized information retrieval platform.
The main contributions of this paper are threefold. (1) A unified framework for the user privacy protection of personalized information retrieval, which has good practical usability. (2) A privacy model for the user privacy protection of personalized information retrieval, which formulates the constraints that should be satisfied for the effective protection of preference privacy, location privacy and query privacy. (3) An implementation algorithm for the privacy model under the framework, and can comprehensively improve the security of all kinds of user privacy on the untrusted server. Overall, this paper presents an important and valuable study attempt to the protection of user privacy in personalized information retrieval, and its study result is of a positive influence on the problem of privacy and security in the field of organizational and end user computing.