To protect the privacy data stored in mobile devices, a secure storage and self-destruction scheme for mobile devices was proposed, which was based on data compression, threshold secret sharing and mobile social networks. In this security scheme, the private data was first compressed with a lossless compression technique, and then the compressed data was encrypted symmetrically by a symmetric key to obtain primitive ciphertext, which was divided into two parts of ciphertext. With time attribute, one part of ciphertext was encapsulated into the Mobile Data Self-destructing Object (MDSO), which was stored in cloud servers. Furthermore, with the symmetric key and time attribute, the other part of ciphertext was processed by the Lagrange polynomial, and the mixture ciphertext shares were generated. At last, these mixture ciphertext shares were embedded into the pictures sharing to social networks. When the authorization was expired, no one could obtain ciphertext block to recombine the original ciphertext, so the security of the privacy data could be protected. The experimental results show that, the sum of the compression and encryption time is only 22 ms when the size of file is 10 KB, which explains the proposed scheme has low performance overhead; furthermore, the results of the comprehensive analysis indicate that the proposed scheme has high security, and it can resist against attacks effectively and protect the mobile privacy data.

Concerning gender tendency hidden in microblog messages posted by microblog users, a novel approach based on rough set theory was proposed to identify microblog user gender. In the proposed approach, a new Representation Model based on Tolerance Rough Set (TRSRM) was devised, which can effectively represent gender characteristics of microblog messages. The experimental results show that the accuracy rate of the proposed approach is 7% higher than frequency model approach by testing messages of 1000 real microblog users, and so the TRSRM achieves better recognition performance.

To deal with the increasingly serious situation of document's security and better protect the controlled documents, in this paper, an identity-based On-The-Fly Encryption (OTFE) and decryption scheme was proposed for the controlled documents, which combined an Identity-Based Encryption (IBE) algorithm with an on-the-fly encryption technique. In the scheme, file system filter driver technology was used to monitor program's behaviors on the controlled documents; meanwhile, the IBE algorithm was used to encrypt and decrypt the controlled documents. Specifically, a new algorithm that associated the original ciphertext and divided the associated ciphertext into two parts stored in different locations was proposed. Therefore, it is impossible for an adversary to obtain the whole ciphertext and further recover the original plaintext. Finally, an elaborate description was made on the scheme from system level and algorithm level. The security analysis indicates that the proposed scheme is able to effectively protect the controlled documents.