Predicting the Political Orientation of Facebook Users Based on Machine Learning Approach

• Main Authors: Cheng-Hua Lin, 林政樺
• Other Authors: 林敏勝
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/j2t4j7

碩士 === 國立臺北科技大學 === 電機工程系 === 106 === Online social networks have become more and more popular in recent years. They provide new platforms for users to exchange and share their information. This has resulted in rich research on predicting the personality traits of individuals from social media. This thesis proposes machine learning-based approaches to predict the users political orientation according to what fan pages theyve liked on Facebook. The experimental results show that (1) the prediction of political orientation on male dataset is more accurate than that on female dataset; (2) the traditional machine learning models perform well even with a small training dataset; (3) the deep learning models need to use a large dataset to get good prediction results.