Analysis On Instant Messaging Protocol Feature And Communication Relationship

Instant messaging (IM) has received much attention among Internet users communicatingeach other, and the IM protocol which characterizes the user status and multi-type data arebecoming diversified and personalized. Presently, in the research literature regarding theprivatized IM applications, the main challenge is about data recognition and abstraction due tothe shortage of publicly available references. In addition, the communication network in theapplication of IM is also an important topic for social networking research.To address the above issues, in this thesis, we study the following aspects of theidentification and applications of IM data in the basis of the mainstream IM protocols.Firstly, as for the private IM protocols, we study on the data packages. We analyze andsummarize the function, the process, data interaction and the message of the mainstream IMprotocols. We also propose two classes of messages, structure-based and text-based, for IM dataanalysis and develop a feasible strategy for data utilization.Secondly, as for the difficulty in IM data recognition and information abstraction, we studythe characteristic identification of protocols based on application data. We develop twoalgorithms drawing on the ideas of Frequency Pattern and Longest Common Subsequence,which can automatically pick up the birthmark and fingerprint features in application layer data.Our experiment shows that the features generated by the proposed algorithms not only do a goodwork in IM protocol identification but also can identify specific types of messages. It is animportant preparing for next research on connected information abstraction from the IM data.Thirdly, as for the social network in the IM data, we study the position and the form of IMdata and choose ways in connected features recognition and abstraction. Combined with theassociate information, we connect the behavior and the user in a communication. Based on thethe partly conversation reconstruction, we get a communication network of IM users accordingto the dynamic transferred IM data, and analyze the relationships in it with social networktechnologies.Finally, as for the issues affecting the relationship strength of IM users, we research on themethodology in relationship evaluation in IM communication strength. We examine thedifferences between the IM networks and other forms of online exchange, propose three aspectsof closely relationship in the communication networks, and study the configuration of specificparameters. Specifically, the method quantify different types of IM applications based on theSocial Presence theory, and the comprehensive evaluation of multiple metrics of userconnectedness based on the TOPSIS theory. Our experiment shows that, comprehensiveevaluation outperforms any single-metric evaluation. The results are useful for the weightednetworks which mostly rely on the communication frequency and the network topology-basedweighted computing. Our study is also helpful for academic research and practice of relevantsocial network analysis.