Cosine similarity-based clustering and dynamic reputation trust aware key generation scheme for trusted communication on social networking

Social networking sites (SNSs) make it possible to connect people and they can communicate with others. Due to the lack of privacy mechanisms, the users in SNSs are vulnerable to some kinds of attacks. Security and privacy issues have become critically important with the fast expansion of SNSs. Most network applications such as pervasive computing, grid computing and P2P networks can be viewed as multi-agent systems which are open, anonymous and dynamic in nature. Moreover, most of the existing reputation trust models (RTMs) do not depend on any clustering structures. The clustering structures are used to effectively calculate the trustworthiness of the network nodes. In this paper, a novel cosine similarity-based clustering and dynamic reputation trust aware key generation (CSBC-DRT) scheme is proposed. For better faced clustering, a cosine similarity measure is estimated for all the nodes on the network. Based on the similarity measure among the nodes, the network nodes are clustered into disjoint groups. The RTM is built in this proposed scheme. Here, an improved MD5 algorithm is explored for key generation and key verification. After the key verification, the trusted measures such as reputation value, positive edge and negative edge values are computed to formulate the trusted network. The proposed scheme performs better than the existing RTM, which provides trusted communication in social networks.