中国房地产价格的研究—基于ε-TSVR模型和VAR模型

Research on the Price of Real Estate in China—Based on ε-TSVR Model and VAR Model

摘要: 中国的房地产价格的有效预测一直是民生的热点问题，从而备受国内外学者的高度关注。本文选用全国2005~2013年的月度数据，在研究中国房地产价格基础上，采用向量自回归(VAR)模型和支持向量回归(ε-TSVR)模型，分别对中国房地产价格进行预测，并比较。研究结论表明：ε-TSVR模型的平均绝对误差MAE、平均绝对百分比误差MPE、均方根误差RMSE值都小于VAR模型，说明ε-TSVR模型对中国房地产价格的预测效果更佳，在房价的预测中有较强的科学性和可行性。

Abstract: China’s price of real estate forecasts has been a hot livelihood issue, and scholars have paid much attention to it. In this paper, the monthly data of the national 2005-2013 years, in the study of Chi-na’s real estate prices, are based on the use of vector auto regression VAR model and support vector regression (ε-TSVR) model, respectively to predict and compare the Chinese real estate prices. The results show that the average absolute error (MAE), the average absolute percentage error (MPE), the root mean square error (RMSE) value of the of ε-TSVR model are less than VAR model, which shows that theε-TSVR model has better forecasting effects on the real estate prices in China.