大数据驱动的我国新能源汽车需求分析  

Big data driven demand analysis of new energy vehicles

作　　者：谭涛 黄泽涛 林雁玲 毕桂灿[2] Tan Tao;Huang Zetao;Lin Yanling;Bi Guichan(Institute of New Rural Development,South China Agricultural University,Guangzhou 510642,China;Institute of New Energy and New Materials,South China Agricultural University,Guangzhou 510642,China)

机构地区：[1]华南农业大学新农村发展研究院,广东广州510642 [2]华南农业大学新能源与新材料研究所,广东广州510642

出　　处：《可再生能源》2020年第7期967-971,共5页Renewable Energy Resources

基　　金：广东省重点研发项目(2019B110209003);国家重点研发项目(2019YFB1503805)。

摘　　要：文章基于网络搜索大数据,以新能源汽车为例,结合统计学和计量经济学理论与方法,利用斯皮尔曼相关系数、协整检验和格兰杰因果关系,检验分析了搜索指数与新能源汽车实际需求之间的关系。以新能源汽车历史销量作为单一变量建立自回归滑动平均模型(ARMA),并与加入了搜索指数的向量自回归模型(VAR)进行比较。结果表明,加入搜索指数的预测模型相较传统的预测模型,在样本期内和样本期外的预测精度分别提高了11.69%和14.95%。该模型只需利用前4个月的新能源汽车销售数据和网络搜索大数据,就能够准确地预测下一个月的需求,在提高预测时效性的同时,也为个人、企业和政府决策提供可靠的依据。As the sales of new energy vehicles continue to rise,the use of big data to analyze and forecast the demand of new energy vehicles will help the industrial development and its supporting industry chain.This research based on online big data,took new energy vehicles as an example,combined statistical and econometric theories,used the Spearman correlation coefficient,Cointegration test and Granger causality test to analyze the relationship between search indexes and actual demand of new energy vehicles.After that,the autoregressive moving average model(ARMA)was established by using the historical sales of new energy vehicles as a single variable and compared with the vector auto-regression model(VAR)which was implemented the search indexes.The results indicated that our model has improved the prediction accuracy during the sample period by 11.69%compared with the traditional prediction model,whose prediction accuracy has increased by 14.95%outside the sample period.This model allowed to use the sales data just from the previous four months and the correlated search data to predict the demand for the next month.It improved the timeliness of forecasting and provided a more accurate basis for individuals,enterprises and governments on decision-making.

关 键 词：新能源汽车 大数据 搜索指数 需求预测 

分 类 号：G203[文化科学—传播学] F713.54[经济管理—市场营销]