本研究的問題在於如何有效地探討社交媒體網站人際網絡感染力與人際結構中的部分權力連結關係？以及如何擷取完整資料以利於分析？透過分析網路社群感染力探討臉書意見領袖與使用行為分析，建立擷取完整資料的新方法。研究方法第一階段以自撰程式擷取資料比對針對log-in 以及Data Mining 找到人際關聯；第二階段透過嵌入智慧型手機資料擷取程式，透過長時期與每日的資料撈取進行分析使用者行為。本研究的主要的重要性在於開發一程式代理人解決網絡分析上工具的不足；分析社群行為方法上有資訊科技(Information Technology, IT)的涉入，分析的結果可以更貼近使用者行為，進而可以得到社群網絡使用者的行為涵義。 This study aimed to explore how to efficiently investigate the relationship between the infectiousness of interpersonal networks in a social media website and the power in an interpersonal structure and how to obtain complete data for analyses. By analyzing the infectiousness of online communities, this study explored the issues of opinion leaders and user behaviors on Facebook and designed a new method to obtain complete data. In the first stage of the research method, a self-developed program was applied to retrieve data and find interpersonal relationships through comparisons of log-in information and data mining. In the second stage, through a program installed in smart phones to collect data, user behaviors were analyzed based on the daily data collected during a long period of time. The key of this study was to develop a software agent to make up the insufficiencies of internet analysis tools. With the involvement of information technology in the analysis, we were able to get a more accurate picture of user behaviors. This was a way to have a further insight into social network users.