No abstract

As the mobile applications become increasing popular, people are installing more and more Apps on their smart phones. In this paper, we answer the question whether it is feasible to predict which App the user will open. The ability for such prediction can help pre-loading the right Apps to the memory for faster execution or help floating the desired Apps to the home screen for quicker launch. We explored a variety of contextual information, such as last used App, time, location, and the user profile, to predict the user's App usage using the MDC dataset. We present three findings from our studies. First, the contextual information can be used to learn the pattern of user's App usage and to predict App usage effectively. Second, for the MDC dataset, the correlation between sequentially used Apps has a strong contribution to the prediction accuracy. Lastly, the linear model is more effective than the Bayesian model to combine all contextual information and for such predictions.

With the continued advances of Web 2.0, health-centered Online Social Networks (OSNs) are emerging to provide knowledge and support for those interested in managing their own health. Despite the success of the OSNs for better connecting the users through sharing statuses, photos, blogs, and so on, it is unclear how the users are willing to share health related information and whether these specialpurpose OSNs can actually change the users' health behaviors to become more healthy.This paper provides an empirical analysis of a health OSN, which allows its users to record their foods and exercises, to track their diet progress towards weight-change goals, and to socialize and group with each other for community support. Based on about five month data collected from more than 107,000 users, we studied their weigh-in behaviors and tracked their weight-change progress. We found that the users' weight changes correlated positively with the number of their friends and their friends' weight-change performance. We also show that the users' weight changes have rippling effects in the OSN due to the social influence. The strength of such online influence and its propagation distance appear to be greater than those in the real-world social network. To the best of our knowledge, this is the first detailed study of a large-scale modern health OSN.

According to actual needs, generalized signcryption scheme can flexibly work as an encryption scheme, a signature scheme or a signcryption scheme. In this paper, firstly, we give a security model for identity based generalized signcryption which is more complete than existing model. Secondly, we propose an identity based generalized signcryption scheme. Thirdly, we give the security proof of the new scheme in this complete model. Comparing with existing identity based generalized signcryption, the new scheme has less implementation complexity. Moreover, the new scheme has comparable computation complexity with the existing normal signcryption schemes.

No abstract