基于主成分分析的我国房地产业周期波动研究

运用主成分分析法构造房地产业景气综合指标（主成分），第一主成分贡献率为0．96829，说明它保留了原始变量96．829％的信息，在房地产业周期波动分析中就可以把其他主成分舍弃。第一主成分与原始变量的相关系数p称为因子负荷量为0．70632，表明第一主成分反映了商品房销售额指标70．632％的信息。选择商品房销售额指标作为代表，来研究房地产业周期变动的特征和规律性是可行的。运用时间序列加法模型和乘法模型分析了我国房地产业周期波动的特征和规律性，我国房地产业循环波动的周期为12-13年。

China real estate periodic fluctuation based on principal component analysis

The real estate economy indicators were constructed by principal component analysis. The contribution rate by first principal component is 0.96829 which means it preserve 96.829% information of original variables. Other principal components can be abandoned in the analysis of real estate periodic fluctuation. The correlation coefficient p of first principal component and original variable is called factor loading, which is 0.70632. It means the first principal component reflects 70.632% information of real estate sales indicators. It is feasible to choose real estate sales indicators to study the characteristics and rules of real estate periodic fluctuation. The time series additive model and multiplicative model were used for the analysis of characteristics and rules The results show that the period of China real estate fluctuation is twelve to thirteen years.