首页|基于图模型分析中国GDP的行业影响因素

基于图模型分析中国GDP的行业影响因素

扫码查看
文章通过《中国统计年鉴》选取近 23 年国内生产总值GDP年度数据,对GDP与九大行业间的关系进行分析,以国内生产总值GDP为被解释变量,九大行业作为解释变量,即农林牧渔业、工业、建筑业、批发和零售业、交通运输仓储和邮政业、住宿和餐饮业、金融业、房地产业、其他.基于偏相关系数选取最佳阈值 0.637 5,并以此构建网络图结构,以GDP与九大行业为节点,相互之间的相依关系为边,进行网络图可视化.根据网络图可得出影响GDP的主要因素有工业、建筑业、金融业、房地产业和其他,其中除房地产业与GDP呈负相关外,其他行业与GDP呈正相关.利用网络图模型分析所得结论与多元统计方法分析所得结论相同,验证了基于偏相关系数与最佳阈值构建的网络图模型分析GDP的影响因素的正确性.但相较于多元统计方法,网络图模型使各节点间的相关关系更为清晰直观,便于理解和分析GDP的行业影响因素.
Analysis on the Industry Influencing Factors of China's GDP Based on the Graph Model
This paper selects annual data of GDP in recent 23 years from China Statistical Yearbook to analyze the relationship between GDP and nine major industries,taking GDP as explained variable and nine major industries as explanatory variable,including agriculture,industry,construction industry,wholesale and retail industry,transportation,post,hotel and catering sectors,finance,real estate,and other sectors.Based on the partial correlation coefficient,the optimal threshold of 0.637 5 is selected,and the network diagram structure is built based on it.The network graph is visualized by taking GDP and nine major industries as nodes and their interdependent relationships as edges.The network graph shows that the main factors influencing GDP are industry,construction,finance,real estate,and other sectors,and only real estate shows a negative correlation with GDP.The conclusion obtained by using the network graph model is the same as that obtained by the multivariate statistical method,which verifies the correctness of the network graph model based on the partial correlation coefficient and the optimal threshold to analyze the influencing factors of GDP.However,compared with multivariate statistical methods,the network graph model makes the correlation between nodes more clear and intuitive,which is convenient to understand and analyze the industry influencing factors of GDP.

graph modelGDP influencing factorsoptimal thresholdnetwork graph

郭洪菊、索利杰、杨帆

展开 >

南京工业大学 数理科学学院,江苏 南京 211816

图模型 GDP影响因素 最佳阈值 网络图

南京工业大学研究生教育教学改革课题

YJG2413

2024

商业观察

商业观察

ISSN:
年,卷(期):2024.10(23)