複迴歸分析法與不動產實價登錄資料的台灣房價估算模型之研究

A Study of Multiple Regression Analysis and Estate Transactions in Taiwan based on Real Estate Price Data

黃耀賢(Yao-Hsien Huang) ； 林惠敏(Hui-Min Lin) ； 陳奕潔(Yi-Chieh Chen) ； 林永鑫(Yung-Hsin Lin)

《創新與管理》 14卷1期 (2018/01) Pp. 1-25

摘要

近年來房價高漲不跌，想購買一間房屋都需要龐大金額支出，如何選購一間符合需求且價格合理的房屋已成為一個重要的課題。本研究將蒐集之房屋實價登錄資料進行統整，加入可能影響房價之因素，例如：總經濟指數、股票指數、批發及零售業、運輸及倉儲業、資訊及通訊傳播業、金融及保險業、不動產及住宅服務業、公共行政及國防、其他的產業經濟指數等變數，利用三種分類方式將資料進行切割，分別為豪宅與一般房屋、縣市與鄉鎮市區、房屋政策時間點，利用皮爾森分析法找出在各種分類情況下影響房價的真正因素，最後將結果篩選後套入複迴歸模型中，建立更準確的房價預估模型。

關鍵字

實價登錄 ； 皮爾森分析 ； 複迴歸模型

並列摘要

With the rising house prices in recent years, buying a house requires a large amount of money. How to select a house that meets the buyer's needs and that is affordable has been an important issue. With the actual real-estate transaction prices collected in the study, potential factors that might affect house prices were also included, such as macroeconomic indicators, stock market index, wholesale and retail trade, transportation and storage industry, information and communication industry, finance and insurance industry, real estate and residential and service, public administration and defence, and economic indexes from other industries. Three types of classification were adopted for data division, including luxury house and average house, city and township, and timing of the housing policy. Pearson product-moment correlation coefficient was utilized to locate genuine factors that affected house prices in each classification. The results were further analyzed with multiple regression for a more accurate model of house price estimate.

並列關鍵字

actual real-estate transaction prices ； Pearson product-moment correlation coefficient ； multiple regression