基于信息觅食理论的消费者在线评论研究

A Study on Consumers’ Online Reviews Based on Information Foraging Theory

摘要: 随着网络的迅速发展，网络消费已经成为现代人生活中不可或缺的一环，消费者通常会根据网路上的评论来进行购物决定。网络评论是影响消费者购买行为的重要因素，正面评论对用户购买行为有重要影响，而负面评价会使消费者产生强烈的抵触情绪。本项目聚焦于用户在积极搜寻与被动接受线上评论时的行为差异，以及这种差异对消费者购买决策的影响。通过混合研究的方式，将量化和质的研究相结合，本文使用了两种数据采集方式：实验数据和文本挖掘。详细剖析了消费者于不同情境下获取在线评论的行为特征及其影响因素，研究开始借助文献研究法梳理了信息觅食理论、在线评论行为以及消费者决策相关的研究成果，以此为后续研究奠定理论基础，其次运用爬虫收集了消费者主动搜索与被动接收评论的行为数据，剖析了影响因素对购买决策所起的作用。借助Python展开文本挖掘，分析从京东、淘宝等电商平台以及社交媒体平台收集到的在线评论数据，运用BERT模型进行情感分析，把评论分为正面、负面和中性，并提取高频关键词，分析评论内容的关注点。研究结果显示，主动搜索评论的消费者对评论质量有较高标准，且更易受评论的可信度以及评论者背景的影响，而被动接收评论的消费者则更多地受评论情感倾向与传播渠道的影响。情感分析结果说明，消费者对产品特性、品牌声誉以及售后服务等方面呈现出较高关注度，本文为企业如何依靠优化在线评论管理策略，提升消费者满意度和购买转化率给出了理论依据和实践建议。

Abstract: With the rapid development of the internet, online consumption has become an indispensable part of modern life. Consumers typically make purchasing decisions based on online reviews. Online reviews are crucial factors influencing consumers’ buying behavior: positive reviews significantly impact users’ purchase intentions, while negative evaluations can trigger strong resistance among consumers. This project focuses on the behavioral differences between consumers actively searching for and passively receiving online reviews, as well as the impact of such differences on their purchase decisions. Adopting a mixed research approach that combines quantitative and qualitative methods, this study uses two data collection methods: experimental data and text mining. It thoroughly analyzes the behavioral characteristics and influencing factors of consumers acquiring online reviews in different contexts. The research begins by using the literature review method to sort out the research findings related to information foraging theory, online review behavior, and consumer decision-making, laying a theoretical foundation for subsequent studies. Secondly, web crawlers are used to collect behavioral data of consumers actively searching for and passively receiving reviews, exploring the role of influencing factors on purchase decisions. Text mining is conducted with Python to analyze online review data collected from e-commerce platforms (such as JD.com and Taobao) and social media platforms. The BERT model is applied for sentiment analysis, classifying reviews into positive, negative, and neutral categories, and extracting high-frequency keywords to identify the focus of review contents. The research results show that consumers who actively search for reviews have higher standards for review quality and are more susceptible to the credibility of reviews and the background of reviewers. In contrast, consumers who passively receive reviews are more influenced by the emotional orientation of reviews and communication channels. Sentiment analysis results indicate that consumers pay significant attention to product features, brand reputation, after-sales service, and other aspects. This study provides theoretical foundations and practical suggestions for enterprises on optimizing online review management strategies to enhance consumer satisfaction and purchase conversion rates.