SIGNIFICANCE FACTOR ANALYSIS OF SOCIAL MEDIA DATA USING DATA MINING METHODS by YUE JING, Maya Thesis Supervisor: Dr. Simon James Fong (Fong Chi Chiu), Associate Professor Co-Supervisor: Dr. Sofia (Zhuang Yan), Assistant Professor Science in E-Commerce Technology With the development of Internet, more and more Internet users use social media to record and share personal status, opinions, and exchange transactions. The social media is the new online media, which can provide the space and platform for users to publish and broadcast information, communicate with each other and collaborate with other persons and other applications. They are expressed in a specific way, like: Blog、Wiki 、 Podcast、Forum、Social network、Bulletin board system and Weibo. For example: Facebook, Sina Weibo, Renren, Ebay and Taobao etc. Upon obtaining the required services through social media, which several opponents can provide it, the users need to evaluate each service provider is a trust value, also according to the trust value to make a choice. So “TRUST” is a big problem on social media, we can trust the one and the website according to one’s trust value and the trust factors, also we can use the trust actors infer degree of friendship as well as investigating what contribute to customer loyalty in e-commerce on online social networks. So computing trust factors quantitatively and fining which one is the important trust factor are the important tasks in social media analysis, especially for online social media where face-to-face interactions are spared. In the past trust factors were obtained by direct questioning during survey or qualitative estimation, the results are then counted usually by simple frequency iv distribution. Here, in this paper, an alternative method by data mining is given that infers quantitatively the relative importance of each trust factor with respective to the predicted class. We will take one social network: Facebook for example to make an experiment and get the results that which are the significance factors. It is done by Feature Selection algorithms. The advantage of using data mining method over simple statistic is that each factor (or Relative Importance Factor) acts as a predictor variable that foretells how the likelihood of an expected occurrence [1]. In other words, is that mean, a-priori knowledge can be assumed and the relative importance factors can be predicted known without the fact that actually occurred. Such relative importance measures would be very useful for computing the weights of variables that are used in some prediction formula and algorithms. So, the author will give a case study of estimating trust and inducing recommendation in Facebook, also do the significance factor analysis of social media data using data mining methods.