基于泰勒级数神经网络方法的我国居住建筑面积数据分析
Analysis of residential building area data of China based on Taylor-series neural network method
摘要：
通过对我国建筑统计年鉴居住建筑面积相关数据统计范围、统计口径、有效年份分析，确定居住建筑面积数据中年末实有面积、竣工面积、人均住宅面积等主要指标的相关性，分别采用泰勒级数神经网络方法和指数线性趋势外推法分析计算了2001—2014年时间序列下城镇和农村居住建筑面积。在此基础上，分析了GDP对居住建筑面积的影响。
Abstract：
Analyses the floor area data of residential buildings in China statistical yearbook, including the statistical data range, data source, effective statistical year. Determines the correlation of main factors such as total floor area of each year, completed floor area and floor area per capita of residential buildings and so on. Adopts Taylor-series neural network method and exponential linear trend extrapolation method to calculate the time series floor area of urban and rural residential buildings of 2001 to 2014, respectively. On the basis, analyses the influence of GDP on floor area of residential buildings.
Keywords:floorarea,residentialbuilding,timeseries,statisticalyearbook,grossdomesticproduct(GDP)