With the deep integration of mobile, social and location, users share information with location tags and friends labels in mobile social networks, which could result in the disclosure of location privacy of the users ...
详细信息
ISBN:
(纸本)9781728105499;9781728105482
With the deep integration of mobile, social and location, users share information with location tags and friends labels in mobile social networks, which could result in the disclosure of location privacy of the users and their friends. Because of the complexity of social circles and the difference of relation strength, users have different location privacy-preserving requirements for different friends. Existing location privacy protection methods and mobile social network services pay less attention to the co-location information attack called CF attack for short. A co-location privacy-preserving system framework is proposed, the key technologies of core module called location privacy-preserving agent are explored, novel calculation method of relation strength based on bi-direction interaction frequency between user and friends is given, and two location privacy-preserving algorithms as a defence against CF attack are designed in the paper, which are LPPC based on the user's coordination and CCTA based on the co-location information concealment and time adjustment. Experimental tests on real data sets show that algorithms proposed could be suitable for different location privacy-preserving scenarios and reflect the mapping between intimacy and privacy-preserving requirement effectively, and the system framework could be extended to the location privacy-preserving application for any mobile social network services.
Plagiarist detection is a complicate topic in plagiarism detection area. Most existing algorithms can only compare similarity between assignments, but cannot detect plagiarist. This paper designed a homework assessmen...
详细信息
Plagiarist detection is a complicate topic in plagiarism detection area. Most existing algorithms can only compare similarity between assignments, but cannot detect plagiarist. This paper designed a homework assessment model based on the South China University of Technology e-learning plagiarism detection module. By analyzing students' learning behavior data collected from the teaching platform, a ranking of the possibilities of plagiarism in students' work is obtained, which provides the basis for judging the plagiarist. And comparing the determination of plagiarism with the actual investigation results, the accuracy of plagiarist determination using the model is significantly improved. The model has been used on the e-learning platform, which provided an effective way for teachers to evaluate assignments.
On the basis of analyzing the classical algorithms of network clustering and community detecting, a new algorithmic method of discovery of community was brought up, which has different thinking from previous algorithm...
详细信息
暂无评论