自回归积分移动平均
残余物
计算机科学
人工神经网络
能量(信号处理)
度量(数据仓库)
数据挖掘
体积热力学
均方预测误差
预测建模
人工智能
时间序列
机器学习
算法
统计
数学
物理
量子力学
作者
Banghua Yang,Jianjun Liu,Dongning Liu
出处
期刊:Communications in computer and information science
日期:2023-01-01
卷期号:: 516-527
标识
DOI:10.1007/978-981-99-2356-4_41
摘要
An important measure of the development of the new energy vehicle market is the prediction of vehicle sales. It is of great significance to complete the construction of relevant supporting facilities, according to the predicted sales volume for the development of the Chinese new energy vehicle industry. Based on this, this paper proposes a combined model that organically combines a single prediction model. Firstly, the ARIMA model is used to predict the linear information in the sales data, and BP neural network model is used to predict the residual sequence between the previous prediction and the actual value. After that, it adds the prediction results to get the final prediction results of new energy vehicle sales. The results verified with the actual sales data show that the prediction accuracy of the ARIMA-BP Residual Optimization Combination model used in this paper is 85.07%. Compared with the single prediction model and the simple weighted combination prediction model, there are general advantages, which can be used for the actual monthly sales prediction of new energy vehicles.
科研通智能强力驱动
Strongly Powered by AbleSci AI