Construct a Mass Appraisal System Based on XCS - An Example of the Kaohsiung Real Estate Market

• Main Authors: CHEN, YEN-HSUN, 陳彥勳
• Other Authors: CHEN, YU-JANG
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/27rg9e

碩士 === 國立屏東大學 === 資訊管理學系碩士班 === 107 === 　　Since the Ministry of the Interior began to enact the actual price registration system, the government accumulated a lot of available real estate price data in the database, which is providing an open and correct information source for the mass appraisal of real estate. 　　From the actual price registration database, this study collected the real estate transaction data in Kaohsiung City from 2013 to 2018 and applied extended classifier system (XCS) to construct a mass appraisal model. In this study, the proposed model was compared with the commonly hedonic price method to the real estate research. Finally, this research calculated separately the two models’ hit rates which are often used to calculate estimated and actual prices. 　　The results have been shown that the proposed mass appraisal model based on the XCS has a higher accuracy than the hedonic price method in the hit rate 10% and the hit rate 20%. In conclusion, XCS is able to be utilized to construct a mass appraisal model.