摘要
本研究以京东平台沙发类家具的在线评论为数据来源,采用Web Scrapy方法采集了3525条评论数据,通过数据预处理方法提升数据质量,运用LDA主题模型,挖掘出消费者关注的主题:舒适性、材质、款式、空间适配性和价格,通过高频关键词分析明确了消费者的关注点。同时,构建了朴素贝叶斯模型对评论进行分类,准确率达到86%。通过对分类结果准确性进行分析以及对不同类别的讨论,帮助企业深入了解消费者满意度。研究总结了两种模型在电商评论分析中的价值,提出未来可通过拓展数据来源、优化数据处理与模型,探索模型结合的方法为电商企业提供更全面的服务,助力在市场竞争中持续发展。
This study utilized online reviews of sofa furniture on the JD platform as the data source,collecting 3,525 review data via the Web Scrapy method.Data quality was enhanced through data preprocessing techniques.The LDA topic model was employed to mine consumer concerns,which were identified as comfort,material,style,spatial adaptability,and price.High-frequency keyword analysis further clarified these concerns.Additionally,a Naive Bayes model was constructed to classify comments,achieving an accuracy rate of 86%。Analysis of the classification accuracy and discussion of different categories provided insights into consumer satisfaction.The study summarized the value of the two models in e-commerce comment analysis and suggested that future research could explore the combination of models by expanding data sources,optimizing data processing and models,and offering more comprehensive services to e-commerce enterprises,thereby aiding their continued development in market competition.
作者
李扬
赵中元
LI Yang;ZHAO Zhongyuan(College of Furnishing and Industrial Design,Nanjing Forestry University,Nanjing 210037,China)
出处
《家具》
2025年第4期29-33,共5页
Furniture
基金
江苏省高等教育教改研究课题(2023JSJG486)。
