基于PCA - GRNN模型的新能源汽车月度销售量预测研究

为预测新能源汽车的月度销售量,提出了一种基于主成分分析(PCA)和广义回归神经网络(GRNN)相结合的预测模型——PCA - GRNN模型.首先,选取动力电池月份装车量、充电基础设施、电池级碳酸锂平均价格、交通和通信类居民消费价格指数、全国城镇调查失业率、汽车制造业工业生产者出厂价格指数等6个指标作为新能源汽车月度销售量的影响因子; 其次,利用主成分分析方法得到可代表6个影响因子的2个主成分,并利用Matlab神经网络工具箱的GRNN神经网络函数构建了广义回归神经网络模型; 最后,将2020—2022年间27个月度的统计数据分别输入到PCA - GRNN、PCA - BP和PCA - Elman模型中进行预测.结果显示, PCA - GRNN模型预测的新能源汽车月度销售量的平均相对误差(4.00%)低于PCA - BP模型和PCA - Elman模型预测的平均相对误差(分别为4.77%和4.29%),因此PCA - GRNN模型在预测新能源汽车销售量方面具有一定的实用性.

Research on forecast of monthly sales volume of new - energy vehicles based on PCA - GRNN model

To forecast the monthly sales of new - energy vehicles, a prediction model based on the principal component analysis with generalized regression neural network(PCA - GRNN)was proposed.Firstly, six indicators were selected as the influencing factors of monthly sales of new - energy vehicles, such as the monthly load of power batteries, charging infrastructure, the average price of battery - grade lithium carbonate, transportation and communication consumer price index, national urban survey unemployment rate, and industrial producer ex - factory price index of the automobile manufacturing industry.Secondly, two principal components representing most of the information of six influencing factors were obtained by PCA, and a GRNN model was constructed using the GRNN neurnal network function of the Matlab neural network toolbox.Finally, the statistical data of 27 months from 2020 to 2022 were input into PCA - GRNN, PCA - BP(principal component analysis - back propagation)and PCA - Elman models for forecasting, respectively.The results show that the mean relative error of the PCA - GRNN prediction model of monthly new - energy vehicle sales(4.00%)was lower than that of the PCA - BP and PCA - Elman models(4.77% and 4.29%, respectively).Therefore, the PCA - GRNN model is practicability in predicting new - energy vehicle sales.