| Summary: | 碩士 === 國立雲林科技大學 === 資訊管理系碩士班 === 99 === In recent years, the issue of Social Network has been paid much attention so that the words and behavior of Internet users and their relationship have become a valuable subject for investigation. But it was not sufficient enough to analyze and find out users’ behavioral patterns only by questionnaires. This study would combine the tag analyzer provided by Firebug with unstructured distributed system, Hadoop, to put users’ public information of Social Network into the HDFS at local site for the convenience of classifying and calculating tags. From the result of processed data, it would be used to observe Internet users’ behavior such as number of words on the Wall, frequency of playing Facebook App, and so on to figure out their interaction relationship.
This study provided a method of retrieving web page, which differed from the crawler method, and combined with Hadoop well characterized by MapReduce to accelerate the web page download and analysis speed. The observation object was set to the Facebook users who had made the Wall public in Taiwan. And the period of data retrieving was from September 2010 to June 2011. The data of users would be retrieved randomly and analyzed to observe and find out their interaction relationship.
|
|---|
