摘要
本文通过对山西省 1998年至 2002年共 50个月份的月度商品房销售面积 (使用商品房销售面积克服了价格波动的影响 )的实证分析,建立了一个结合普通线性模型 (OLS)、非参数核估计(KernelEstimation)、以及时间序列分析(TimeSeries)的统计预测模型。此模型既能够很好地分离出商品房销售面积随时间的增长趋势,又能够很好地刻画它在各个年度周期内部光滑的季节规律。基于此模型,山西省每年商品房销售面积的规律被刻画为四个周期,即:增长期、衰退期、调整期、以及结束期。从中不难看出商品房销售规律的个性。在分析各地区商品房市场规律时,不能照搬全国商品房市场的相关规律,更不能照搬国外的测算结果。
By using Shanxi province’s commercial apartment monthly sales data from 1998-2000,we build up a forecasting model, which combined the idea of ordinary linear regression model, nonparametric kernel estimation, and time series analysis. The proposed model can well differentiate the long-term sales increase trend while maintain a good representation of the seasonal effect in a smooth manner. Based on such a model, we classify Shanxi commercial apartment yearly sales trend into four different periods. They are, namely,increasing period, degrading period, adjusting period,and ending period,from where we are able to see the specialty of the Shanxi real estate market.
出处
《数理统计与管理》
CSSCI
北大核心
2005年第2期7-13,共7页
Journal of Applied Statistics and Management