Department of Business Administration (Master’s Program in Management) Fu Jen Catholic University

• Main Authors: Chun-Wen, Lin, 林君紋
• Other Authors: Dr. Lee-Meh, Lee
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/26980033587841878731

碩士 === 輔仁大學 === 企業管理學系管理學碩士在職專班 === 101 === The purpose of this research is to use three commonly adopted data mining techniques, namely discriminant analysis, logistic regression, and artificial neural networks in building classification models aiming to identify the significant variables which may affect real estate transaction and hence can provide useful information for better resource allocation and customer relationship management. In order to verify the feasibility of the proposed idea, one dataset from a construction company in Taipei was adopted in building classification models. The empirical results indicate that back-propagation neural network has better classification accuracy in comparison with logistic regression and discrimiant analysis under different performance criteria. Besides, the obtained significant variables can also provide useful information for better customer care and relationship management.