Prediction of House Price Index in Taipei

• Main Authors: Chih-Pin Lai, 賴智彬
• Other Authors: Ping-Chen Lin
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/6w5p3s

碩士 === 國立高雄應用科技大學 === 金融資訊研究所 === 102 === Domestic prices have been high in recent years, April 2013, Mr. Minister of Finance Mr. Zhang Shenghu Taipei Vice Mayor Zhang Jin and held a press conference and hoped to lower 30% of real estate price by tax. It refers that the real estate market is very hot now. Discard of government intervention, the purpose of the study is whether we can use the price index correlation with other macroeconomic variables to predict future price movements and the factors to affect the fluctuation of real estate market in Taiwan, in order to provide reference for the academic, commercial purposes, or to provide the first purchase group, first repurchase group the awareness of price change, and to provide objective indicators for real estate investors and researchers to make responses. In this study, we use 106 data, from January 2005 to November 2013,to predict future house price index in greater Taipei area and analyze the data by a random walk(RW), multiple regression analysis(MR), stepwise regression analysis (SW), autoregressive integrated moving average model (ARIMA model) time series method. And analyze the data by mean squared error (MSE), sum of square for error (SSE) and absolute error (MAE), three kinds of assessment criteria, Performance comparison of different methods to predict the performance. Then compare the results of different methodologies to accurately predict changes in the 2nd hand real estate market in Taiwan. The empirical results show that in terms of predicting the performance, a autoregressive integrated moving average model is the best way to predict the performance, and the next one is random walk,and then is Multiple regression analysis.Stepwise regression analysis is the worst. In terms of prediction accuracy, Multiple regression analysi prediction accuracy was highest, Random Walk prediction accuracy rate ranked the next, Stepwise regression analysis and autoregressive integrated moving average modeand other two prediction accuracy was the worst. In terms of correlation between macroeconomic variables and the price index, Taiwan's five major banks home loan interest rates, Consumer price index, Construction Cost Index, Consumer price index housing rental category, the unemployment rate, TAIEX Index and the monthly average price of international crude oil in Dubai, these seven variables and overall economic relevance price index are more significant and the correlation are higher.