摘要
针对汽车销量时间序列数据的季节性、非线性性、非平稳性等复杂特征,提出一种融合变分模态分解(Variational Mode Decomposition,VMD)的卷积神经网络(Convolutional Neural Network,CNN)、门控循环单元(Gated Recurrent Unit,GRU)和长短期记忆(Long Short-Term Memory,LSTM)神经网络组合的汽车销量预测方法,通过VMD将汽车销量时序数据进行分解,利用CNN提取关键特征,并通过GRU与LSTM捕捉汽车时序数据的时间依赖关系.实验表明该方法有较好的预测性能.
To address the complex characteristics about time series data of automobile sales,such as seasonality,non-linearity,and non-stationarity,this study proposed a forecasting model that integrated Variational Mode Decomposition(VMD)with Conv-olutional Neural Network(CNN),Gated Recurrent Units(GRU),and Long Short-Term Memory(LSTM)networks.The proposed method began by decomposing the automobile sales data using VMD,followed by the extraction of key features through the appli-cation of CNNs.Subsequently,the model employed GRU and LSTM networks to capture the temporal dependencies inherent in automobile sales data.Experimental results demonstrated that the proposed approach exhibited superior forecasting performance.
作者
范亚茹
向兵
FAN Ya-ru;XIANG Bing(School of Mathematics,Southwest Minzu University,Chengdu 610041,China)
出处
《西南民族大学学报（自然科学版）》
CAS
2024年第4期441-446,共6页
Journal of Southwest Minzu University(Natural Science Edition)
关键词
汽车销量预测
长短期记忆网络
卷积神经网络
组合预测
automobile sales forecasting
long short-term memory network
convolutional neural network
composite forecasting
