融合CNN和WDF模型的电商企业商品销量预测研究

doi:10.3778/j.issn.1002-8331.2312-0318

摘要/Abstract

摘要： 为了适应电商企业商品销量数据规模大、维度高、非线性等特征，并提高销量预测的准确性，创新性地提出一种卷积神经网络融合加权深度森林（CNN-WDF）的销量预测方法。利用卷积神经网络（CNN）处理高维数据的优势对电商企业商品销量数据进行特征提取，降低冗余度和模型训练复杂度。提出一种改进的加权深度森林模型（WDF）进行商品销量预测。该模型依据各个子树的预测准确率计算每一级森林中该子树的权重以提高整体预测准确性，且相对于传统深度网络模型具有超参数少、可解释性强等优点。利用京东商品销量数据进行实验验证，结果表明： CNN-WDF融合模型在不同规模京东销售数据集上，预测准确率均显著高于其他对比模型，且随着数据集规模的扩大，预测准确率提高更加明显。

关键词: 商品销量预测, 深度学习, 融合模型, 卷积神经网络, 加权深度森林

Abstract: To enhance the accuracy of sales prediction for e-commerce enterprises’ merchandise sales data, which often exhibit large scale, high dimensionality, and non-linearity, the paper proposes a novel sales prediction method called convolutional neural network integrated weighted deep forest (CNN-WDF). Firstly, a convolution neural network (CNN), which has the advantage of handling high-dimensional data, is used to extract features of high-dimensional data to reduce redundancy and model training complexity. Secondly, an improved weighted deep forest model (WDF) is proposed for commodity sales prediction, in which the weight of each subtree in each forest level is calculated based on its prediction accuracy. The model can not only improve the overall prediction accuracy, but also has the advantages of fewer hyperparameters and stronger interpretability compared with traditional deep network models. The experimental results using Jingdong commodity sales data show that the prediction accuracy of the CNN-WDF fusion model is significantly higher than that of other comparative models on the same datasets of different sizes. Moreover, the prediction accuracy of this model improves even more with the increase in dataset size.

Key words: commodity sales forecast, deep learning, integrated model, convolutional neural network, weighted deep forest

袁瑞萍, 魏辉, 傅之家, 李俊韬. 融合CNN和WDF模型的电商企业商品销量预测研究[J]. 计算机工程与应用, 2025, 61(2): 335-343.

YUAN Ruiping, WEI Hui, FU Zhijia, LI Juntao. Research on Commodity Sales Forecast of E-Commerce Enterprises Integrating CNN and WDF Model[J]. Computer Engineering and Applications, 2025, 61(2): 335-343.

参考文献

[1] 李杰, 王玉霞, 赵旭东. 电商企业商品销量的预测方法[J]. 统计与决策, 2018, 34(22): 176-179.
LI J, WANG Y X, ZHAO X D. Forecast method of commodity sales of E-commerce enterprises[J]. Statistics & Decision, 2018, 34(22): 176-179.
[2] 石通斌. 基于Xgboost的商品需求预测与分仓规划[D]. 成都: 西南财经大学, 2017.
SHI T B. Commodity demand forecasting and warehouse planning base on Xgboost model[D]. Chengdu: Southwestern University of Finance and Economics, 2017.
[3] 刘丽, 裴行智, 雷雪梅. 基于时间卷积注意力网络的剩余寿命预测方法[J]. 计算机集成制造系统, 2022, 28(8): 2375-2386.
LIU L, PEI X Z, LEI X M. Temporal convolutional attention network for remaining useful life estimation[J]. Computer Integrated Manufacturing Systems, 2022, 28(8): 2375-2386.
[4] 刘吉华, 张梦迪, 彭红霞, 等. 基于卷积神经网络的汽车销量预测模型[J]. 计算机科学, 2021, 48(1): 178-183.
LIU J H, ZHANG M D, PENG H X, et al. Automobile sales forecasting model based on convolutional neural network[J]. Computer Science, 2021, 48(1): 178-183.
[5] 荣飞琼, 郭梦飞. 基于卷积神经网络的在线产品销量预测分析研究[J]. 西北民族大学学报 (哲学社会科学版), 2019, 40(2): 15-26.
RONG F Q, GUO M F. On suitability of online product sales prediction model based on convolutional neural networks[J]. Journal of Northwest minzu University (Philosophy and Social Sciences), 2019, 40(2): 15-26.
[6] 邱俊杰, 郑红, 程云辉. 基于多尺度LSTM预测模型研究[J]. 系统仿真学报, 2022, 34(7): 1593-1604.
QIU J J, ZHENG H, CHENG Y H. Research on prediction of model based on multi-scale LSTM[J]. Journal of System Simulation, 2022, 34(7): 1593-1604.
[7] 王逸文, 王维莉. 基于LSTM-RELM组合模型的电商GMV预测研究[J]. 计算机工程与应用, 2023, 59(10): 321-327.
WANG Y W, WANG W L. Research on GMV prediction of e-commerce based on LSTM-RELM combination model [J]. Computer Engineering and Applications, 2023, 59 (10): 321-327.
[8] 胡越, 罗东阳, 花奎, 等. 关于深度学习的综述与讨论[J]. 智能系统学报, 2019, 14(1): 1-19.
HU Y, LUO D Y, HUA K, et al. Overview on deep learning[J]. CAAI Transactions on Intelligent Systems, 2019, 14(1): 1-19.
[9] 高学金, 孟令军, 高慧慧. 基于注意力LSTM的多阶段发酵过程集成质量预测[J]. 控制与决策, 2022, 37(3): 616-624.
GAO X J, MENG L J, GAO H H. Integrated quality prediction of multi-stage fermentation process with attention-based LSTM[J]. Control and Decision, 2022, 37 (3): 616-624.
[10] 谢坤, 容钰添, 胡奉平, 等. 基于数据集成的随机森林算法[J]. 计算机工程, 2020, 46(12): 290-298.
XIE K, RONG Y T, HU F P, et al. Random forest algorithm based on data integration[J]. Computer Engineering, 2020, 46(12): 290-298.
[11] ZHOU Z H, FENG J. Deep forest: towards an alternative to deep neural networks[C]//Proceedings of the 26th International Joint Conference on Artificial Intelligence, 2017: 3553-3559.
[12] GUO Y, LIU S H, LI Z H, et al. BCDForest: a boosting cascade deep forest model towards the classification of cancer subtypes based on gene expression data[J]. BMC Bioinformatics, 2018, 19(5): 1-13.
[13] SU R, LIU X Y, WEI L Y, et al. Deep-resp-forest: a deep forest model to predict anti-cancer drug response[J]. Methods, 2019, 166: 91-102.
[14] CHU Y Y, KAUSHIK A C, WANG X G, et al. DTI-CDF: a cascade deep forest model towards the prediction of drug-target interactions based on hybrid features[J]. Briefings in Bioinformatics, 2021, 22(1): 451-462.
[15] ZHOU M, ZENG X H, CHEN A Z. Deep forest hashing for image retrieval[J]. Pattern Recognition, 2019, 95: 114-127.
[16] LECUN Y, BENGIO Y, HINTON G. Deep learning[J]. Nature, 2015, 521: 436-444.
[17] 葛绍林, 叶剑, 何明祥. 基于深度森林的用户购买行为预测模型[J]. 计算机科学, 2019, 46(9): 190-194.
GE S L, YE J, HE M X. Prediction model of user purchase behavior based on deep forest[J]. Computer Science, 2019, 46(9): 190-194.
[18] 宫振华, 王嘉宁, 苏翀. 一种加权的深度森林算法[J]. 计算机应用与软件, 2019, 36(2): 274-278.
GONG Z H, WANG J N, SU C. A weighted deep forest algorithm[J]. Computer Applications and Software, 2019, 36(2): 274-278.
[19] 陈寅栋, 李朝锋, 桑庆兵. 卷积神经网络结合深度森林的无参考图像质量评价[J]. 激光与光电子学进展, 2019, 56 (11): 131-137.
CHEN Y D, LI C F, SANG Q B. Unreferenced image quality evaluation based on convolutional neural network combined with deep forest[J]. Laser & Optoelectronics Progress, 2019, 56(11): 131-137.
[20] 陈彦如, 张涂静娃, 杜千, 等. 基于深度森林的高铁站室内热舒适度等级预测[J]. 计算机应用, 2021, 41(1): 258-264.
CHEN Y R, ZHANG T J W, DU Q, et al. Prediction of indoor thermal comfort level of high-speed railway station based on deep forest[J]. Journal of Computer Applications, 2021, 41(1): 258-264.