基于复杂网络技术的电商平台用户情感影响力传播研究

Research on the Dissemination of User Emotional Influence on E-Commerce Platforms Based on Complex Network Technology

摘要: 随着互联网技术的飞速发展，电商平台外卖服务已成为人们日常生活中不可或缺的一部分，用户对于电商平台外卖服务情感态度则能够体现出其认知和判断。本文基于复杂网络技术，对电商平台用户情感影响力的传播进行了深入研究。即通过构建MIGM节点影响力评价模型并使用真实数据集和肯德尔(Kendall)相关系数进行评估，爬取用户转发、评论、点赞、节点影响力MIGM值数据构建权重评价指标，采用卷积神经网络–长短期记忆网络(Convolutional Neural Networks-Long Short Term Memory, CNN-LSTM)对节点情感状态进行划分，从而用于构建复杂网络模型以分析用户情感在复杂网络中的传播路径和模式。研究发现：高影响力用户的情感观点具有更广泛的传播性，且节点密度更为集中，呈现更为显著的情绪“簇团状”结构，中影响力用户则处于复杂网络边缘，情感节点的网络分布相对发散。此外，通过分析不同圈层用户的情感状态，发现所构建的复杂网络中用户对电商平台商家的做法态度持多元化，用户多聚焦于支持商家减少资源浪费的举措，以及秉持关注企业的积极应对措施和信息公开的态度。

Abstract: With the rapid development of Internet technology, the food delivery service on e-commerce platforms has become an indispensable part of People’s Daily lives. Users’ emotional attitudes towards the food delivery service on e-commerce platforms can reflect their cognition and judgment. Based on complex network technology, this paper conducts an in-depth study on the dissemination of emotional influence of users on e-commerce platforms. That is, by constructing the MIGM node influence evaluation model and using the real data set and the Kendall correlation coefficient for assessment, the weight evaluation indicators are constructed by crawling the data of user forwarding, comments, likes, and node influence MIGM values. The Convolutional Neural Networks-Long Short Term Memory (CNN-LSTM) is adopted to divide the emotional states of nodes. Thus, it is used to construct complex network models to analyze the propagation paths and patterns of user emotions in complex networks. The research finds that the emotional viewpoints of highly influential users have wider spreadability, and the node density is more concentrated, presenting a more significant emotional “cluster-like” structure. In contrast, moderately influential users are at the edge of complex networks, and the network distribution of emotional nodes is relatively divergent. Furthermore, by analyzing the emotional states of users in different circles, it is found that in the constructed complex network, users’ attitudes towards the practices of e-commerce platform merchants are diversified. Users mostly focus on measures to support merchants in reducing resource waste, as well as positive response measures and an attitude of information disclosure that pay attention to enterprises.