基于统计年鉴的全国及省级居住建筑面积数据分析及驱动因素模型

摘要

基于统计年鉴,采用泰勒级数神经网络方法预测了全国及省级2001-2015年时间序列宏观居住建筑面积数据.研究发现,全国城镇居住建筑面积平均年增幅远高于农村地区,省级居住建筑面积总量与GDP呈正相关,省级城镇居住建筑面积占总居住建筑面积比例与城镇化率呈正相关.选取GDP作为驱动因素,建立了全国及省级居住建筑面积预测模型.结果表明,模型相对偏差较小,验证了模型的合理性.%Based on the statistical yearbook,predicts the national and provincial macroscopic residential building area data from 2001 to 2015 using the Taylor series neural network method.The research shows that the average annual growth rate of urban residential building area is much higher than that of rural area,the total amount of residential building area at provincial level is positively correlated with GDP,and the ratio of urban residential building area to total residential building area is positively correlated with the urbanization rate.Establishes a national and provincial residential building area prediction model when GDP is selected as a driving factor.The result shows the small relative deviation of the model,which verifies the rationality of the model.