当前位置: X-MOL 学术Electron. Commer. Res. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Cross-border e-commerce platform for commodity automatic pricing model based on deep learning
Electronic Commerce Research ( IF 3.462 ) Pub Date : 2020-11-12 , DOI: 10.1007/s10660-020-09449-6
Lina Guo

With the innovation of information technology and the rise of the Internet economy, cross-border e-commerce has grown up to be an important means and strategy for enterprises to seek rapid development. This paper proposes a model that fuses CNN (Convolutional Neural Network) and attention mechanism to encode image features, and selects the image features of commodities. A 5-layer CNN without a fully connected layer is constructed to initially extract image features, and then a set of attention mechanism strategies is designed. This strategy is used to select the image features that have the greatest impact when generating words at different times. Considering the characteristics of quantitative indicators of the pricing model, this paper transforms this evaluation process of consumers into price perception. Corresponding mathematical model is set up to improve and expand the original probability unit model. The consumer selection model is utilized to obtain a prediction of product market share, and a nonlinear constraint programming is established to determine the optimal price. The strategy takes into account the changed market shares of consumer characteristics and product quality evaluation results. In the two-layer hybrid channel supply chain model, retailers and manufacturers all use third-party platforms when they achieve maximum benefits; when price cross-elasticity coefficients and third-party platform usage fees are independent variables of influencing factors, retailers are dispersed on CNN to get the most profit under the pricing strategy. Similarly, when the unit product tax difference is the independent variable of the influencing factors, the manufacturer is also the most profitable under the CNN decentralized pricing strategy.



中文翻译:

基于深度学习的跨境电子商务商品自动定价模型平台

随着信息技术的创新和互联网经济的兴起,跨境电子商务已经成为企业寻求快速发展的重要手段和战略。提出了一种融合CNN(卷积神经网络)和注意力机制对图像特征进行编码的模型,并选择商品的图像特征。构建不具有完全连接层的5层CNN,以最初提取图像特征,然后设计一组注意机制策略。此策略用于选择在不同时间生成单词时影响最大的图像特征。考虑到定价模型定量指标的特点,本文将消费者的评估过程转化为价格感知。建立了相应的数学模型,以改进和扩展原始的概率单位模型。利用消费者选择模型来获得产品市场份额的预测,并建立非线性约束程序来确定最佳价格。该策略考虑了消费者特征和产品质量评估结果变化的市场份额。在两层混合渠道供应链模型中,零售商和制造商在获得最大收益时都使用第三方平台。当价格交叉弹性系数和第三方平台使用费是影响因素的独立变量时,零售商在定价策略下分散在CNN上以获取最大的利润。同样,

更新日期:2020-11-12
down
wechat
bug