Clustering Friends on Facebook Based on Machine Learning Approach

• Main Authors: Zhi-Kai Fan, 范智凱
• Other Authors: 林敏勝
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/5wcy55

碩士 === 國立臺北科技大學 === 電機工程系 === 106 === This thesis presents a study of clustering friends on Facebook based on machine learning methods. In the beginning, we introduce some commonly used traditional methods for clustering friends on social networks; then, we present how to crawl and analyze data from Facebook. Finally, we propose a machine learning-based clustering algorithm and use friends likes data on Facebook as the similarity measurement. We apply so-called classes-to-clusters evaluation to measure the accuracy of the clustering methods. The results of the experiment demonstrate that the proposed method clusters friends on Facebook with higher accuracy than previous methods. Furthermore, we use simulation way to make experiment so as to investigate how the parameters of social networks affect the results of clustering friends on Facebook.