Privacy-preserving Algorithms And Its Applications In MSNS

In recent years, along with the rapid development of the Internet and computer technology, various and feature-rich intelligent mobile terminals, such as smartphones, tablets and other equipments become more and more popular. Foursquare, Loopt and sina weibo and other social network applications especially based on the mobile end service got rapid growth. In terms of domestic, the success of sina weibo shows the huge market value of social networking service. But it also faces many challenges while creating great value. Personal privacy information Leaked because of the use of equipments when people enjoy the services such as location-based mobile social network service becomes one of the core problems, which has received wide attention of scholars. How to protect the privacy of data and prevent sensitive information from leakage and achieve the optimal balance between the quality of service has become the major challenge.In this paper, a mobile social network privacy protection is introduced for sina weibo from the following three aspects: 1) data query: a method is established based on context-aware adaptive position K anonymous privacy protection algorithm, this method is based on the design of a road network density and protect the privacy control operator combined with K anonymous model is used to query the location of the privacy protection serv ice.2) data release level: a diversity of L model based on improved hierarchical clustering, the method using the Gini Index to measure the distribution of sensitive value attribute, and the information loss and property distribution and constrained clustering of merging process for anonymous equivalence class.3) mining level: to design a kind of suitable for mobile environment, a collaborative filtering recommendation algorithm is introduced based on difference of privacy. This method firstly by using Chinese word segmentation and other technical for building user behavior, and then into the user similarity matrix based on local sensitivity mechanism of the upper bound of Laplace satisfy the difference requirements.