Measuring Tie Strength in Online Social Networks

• Main Authors: Chien-Chou Chen, 陳建州
• Other Authors: Yuh-Jzer Joung
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/01096939793373698908

碩士 === 臺灣大學 === 資訊管理學研究所 === 98 === In the previous studies, people have found many properties of social network, such as “small world phenomenon” and high clustering. But, most of the researches merely consider whether this is a tie, rather than the strength of tie. Since the relationship between people is not identical or symmetric, we regard the network as a directed graph, and the tie on it have its “weight”. Under this circumstance, we wonder what is the role of tie strength in those properties of social network. Therefore we parse data of a blogging site, and record the interaction between the users on it. After defining different types of relationship, we then count the times of action as the strength of tie. First, we examine the distribution of tie strength, from both network and users’ view, and then compare to the distribution of link degree. To evaluate the node, we propose a new measurement, calculated by the strength of tie. Next, we take tie strength into account of average shortest path and cluster coefficient. We find out some interesting characteristics people did not pay attention to before. Finally, we examine the symmetry of the network and analyze the correlation between tie strength of nodes for the first time. We believe that this empirical study of tie strength can help us further understand the macro characteristics of social network, and provide a new perspective for the future researches.