Social Network Prediction Based On Fuzzy System

Social networking refers to a relationship network in which the individuals are bonded by certain relationship. It is a form of social organization based on networks rather than groups. With the rise of industrialization, urbanization and computer science as well as newly carried out communications technology, human society presents the trend of more and more networks. Moreover, with the development of information technology in recent years, social networking is becoming a new interdisciplinary research hotspots.As a method of sociological research, the main research purpose of social network analysis is to study the relationship between individuals in society. Social network prediction as one of the hotspot of social network analysis, due to the constantly varying of the real-world network, the method make a reasonable assessment of the development trend of social networks by analyzing topology and observe the evolution process of social networks. Uncertainty is prevailing in real world network relationships. Traditional network analysis could not properly describe and reflect the real network because of subjective measurement, as the traditional relationship between certainty and probability theory is difficult to truly reflect the objective reality of things. We introduced fuzzy system in this thesis, traditional social network analysis were combined with general fuzzy method intuition fuzzy method separately, transforming traditional deterministic network into fuzzy network: in the process of fuzzilizing the network, several improved basic social networks measures and similarity measure are combined to define general fuzzy network and intuition fuzzy network. In order to reduce the amount of calculation and usage of space, the range of calculation will be extended from a single computing node to communities within the network---obtaining the fuzzy relationship within the network in the network similarity algorithm based on community division. Markov chain are then modified to fit the community-based fuzzy network. Finally, a fuzzy method is represented to predict and analyze the trend of social network evolution and the result are satisfying.Experiment result shows that network with combination of fuzzy is more flexible and accurate on the social network evolution trend forecasting and evaluation, thus have a certain superiority in the social network analysis.