Bilevel Feature Extraction-Based Text Mining for Fault Diagnosis of Railway Systems

Abstract-Among the applications of the internetand cloud computing, online social network (OSN) is a very popular service. Since a lot of personal information is stored on the OSN platform, privacy protection on such an application has become a critical issue.Apart from this, OSN platforms need advertisement revenue to enable continued operations. However, if the users encrypt their messages, then OSN providers cannot generate accurate advertisement to users. Thus, how to achieve both privacy preserving and accurate advertisement is a worth-discussing issue. Unfortunately,none of the works on OSNs can achieve both privacy preserving and accurate advertisement simultaneously. In view of this, we propose the first multireceiver predicate encryption scheme for OSN platforms. Not only does the proposed scheme protects the users’ privacy but it achieves customized advertisement as well.
Compared with other predicate encryptions deployed in OSN platforms, the proposed scheme gains much shorter ciphertext. The semantic security and attribute hiding of the proposed scheme are proved under the standard model