基于AIGC+NLP的电子商务系统——内容生成与智能交互的应用研究

E-Commerce System Based on AIGC+NLP:Application Research of Content Generation and Intelligent Interaction

为解决传统电子商务系统缺乏个性化服务的问题,提出了基于生成式人工智能(artificial intelligence generated content,AIGC)和自然语言处理(natural language processing,NLP)的设计方案,并进行了关键算法和组件的原型实现。首先,利用K均值聚类算法(K-means clustering algorithm,K-means)对用户进行聚类,生成用户画像,进一步借助AIGC和NLP技术创建虚拟主播,为不同类型的用户提供个性化服务。在NLP相关研究中提出的生成对抗网络-双向编码器表征法(generative adversarial networks-bidirectional encoder representations from transformers,GAN-BERT)模型可用于实现虚拟主播与用户的智能对话功能,基于开源数据集上的对比测试,该模型效果相比于其他模型有明显提升,双语替换评测(bilingual evaluation understudy,BLEU)值可达到44.25。电子商务系统采用前后端分离的开发模式,前端引入Vue.js框架对数据进行双向绑定,后端采用Spring Cloud架构和微服务组件,保证服务的可靠性和稳定性。本系统的设计与实现过程将为其他电商平台的开发和优化提供不错的参考价值,进一步推动电商平台的发展。

In order to solve the problem of lack of personalized services in traditional e-commerce systems,a design scheme based on artificial intelligence generated content(AIGC)and natural language processing(NLP)is proposed,and the key algorithms and components are prototyped.Firstly,the K-means clustering algorithm(K-means)is used to cluster users,generate user portraits,and virtual anchors are further created with the help of AIGC and NLP technology,and personalized services are provided for different types of users.The generative adversarial networks-bidirectional encoder representations from transformers(GAN-BERT)model proposed in NLP can be used to realize the intelligent dialogue function between virtual anchors and users,and based on the comparative test on the open source dataset,the effect of the model is significantly improved compared with other models,and the bilingual evaluation understudy(BLUE)value can reach 44.25.The e-commerce system adopts the development mode of front-end and back-end separation,the front-end introduces a Vue.js framework to bind data in both directions,and the back-end adopts Spring Cloud architecture and microservice components to ensure the reliability and stability of the service.The design and implementation process of this system will provide a good reference value for the development and optimization of other e-commerce platforms,and further promote the development of e-commerce platforms.