Machine learning-based system for asset price prediction a case study on Taiwanese estate price

• Main Authors: Jing-Tzung Chen, 陳敬宗
• Other Authors: Chia-Mu Yu
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/28qf65

碩士 === 元智大學 === 資訊工程學系 === 104 === Nowadays, real estate prices are becoming unstable in Taiwan, because of the various rumor, media, and information on the Internet. Thus, this thesis is aimed to develop an automatic system for analyzing a wide range of information on the Internet and predicting the trend of real estate prices. In our system, Latent Dirichlet Allocation (LDA) is used to categorize various webpages based on the internet topic-model. After that, we predicted and analyzed price of assets in Taiwan through semi-supervised learning and supervised learning skills, as well as providing considerable referential information. Our preliminary results confirms the accuracy of our proposed system.