Summary: | 碩士 === 康寧大學 === 資產管理與城市規劃研究所 === 100 === Service industry is the main force of the economic structure in Tainan City. In location, it relatively concentrates in central and western area, east area, and Yongkang area, next, in north area and south area. The five areas are the gather places of business circles or the settlements with strong business phenomenon. Economy in these areas is very active. The changes of real estate is obviously related with the gather places of business circles and the settlements with strong business phenomenon. Since Dec. 25, 2009, original Tainan county and city merged and upgraded to be a directly-controlled municipality. To know the distribution trend and influencing factors of housing price in Tainan City region, the study mainly focuses on Rende area and Guiren area, its collection method adopts Ministry of Internal Affairs Land Administration Department World Wide Web- real estate transaction price data, it discusses the change trend of real estate from the first quarter of 2005 to the first quarter of 2012 to find out price change factors of real estate in Tainan City Rende and Guiren areas and set up models. Through the measurement method of regression analysis, price change factors in Tainan City Rende and Guiren areas are discussed, which can be used as the reference of purchasing second-hand houses.
The study uses three kinds of regression models and found out that linear model 1 is the best prediction model, its regression equation R2 is 0.783, influencing factors are location, road width, house age, land area (m2), floor area (m2), unemployment rate (%), consumer price index, rent and annual growth rate; indistinctive factors are street frontage relation and structure relation. According to empirical results, the prediction based on data from the first quarter of 2005 to the fourth quarter of 2011 shows that error rate is within 20% , accuracy rate reaches 68.07%, MAPE is 16.94. when the first quarter of 2012 is adopted to make prediction, error rate is within 20%, accuracy rate reaches 85.25%, MAPE is 13.55. From 2005 to 2011 in Rende area, error rate is within 20%, accuracy rate reaches 71.5% , MAPE is 18.588; From 2005 to 2011 in Guiren area, error rate is within 20%, accuracy rate reaches 64.8%, MAPE is 15.177. For 2012 Rende area prediction, error rate is within 20%, accuracy rate reaches 85.3% , MAPE is 13.924; For 2012 Guiren area prediction, error rate is within 20%, accuracy rate reaches 89.5%, MAPE is 12.393.
|