高级检索

基于文本挖掘的高跟鞋在线评论分析与关注要素研究

Analysis of Online Reviews of High Heel Shoes and Research on Attention Factors based on Text Mining

  • 摘要:
    目的 为帮助商家和企业精准把握消费者在线购买高跟鞋时的关注要素。
    方法 爬取天猫平台38 199条高跟鞋商品评论,采用词频分析、关键词共现分析、情感分析和聚类分析对评论数据展开文本挖掘。
    结果 词频与共现分析表明,“好看”和“舒服”是消费者的核心关注点;情感分析显示,正面评论多于负面评论,综合评价中“正-负”评论占比小(1.02%),却暴露出产品的隐性问题;聚类分析表明,负面评论聚焦于外在品质、穿着不适、服务不满、质量缺陷和尺码不准。
    结论 外观和舒适性是消费者对高跟鞋产品的核心关注要素;当前消费者对产品外观表示认可,但在舒适性、质量和尺码适配方面仍有较大改进空间;追加评论及其情感转变内容是发现产品潜在问题的重要来源。

     

    Abstract:
    Objective With the rising demand for high heel shoes and the prevalence of online shopping as the mainstream consumption method, there is a large amount of unexplored consumer preference and pain point information hidden in online reviews of high heel shoes. This study aims to analyze online high heel shoes reviews to identify the key factors that consumers consider in their purchasing decisions, providing data support for enterprises and merchants to enhance product competitiveness and customer satisfaction.
    Method The study collected 38,199 valid reviews of the top 200 best-selling high heel shoes on Tmall. Text mining methods were comprehensively applied: the Term Frequency-Inverse Document Frequency algorithm was used to extract high-frequency keywords and create word clouds; the KH Coder was used to construct a keyword co-occurrence network; the SnowNLP library was used for sentiment classification to identify positive and negative reviews as well as sentiment transition types; and the LDA topic model was used to cluster positive and negative reviews separately, generating five themes for each and obtaining their corresponding keyword lists.
    Result Word frequency and co-occurrence analysis indicated that "good-looking" and "comfortable" were the two most concerned dimensions by consumers, which together formed the core basis of satisfaction. Sentiment analysis showed that approximately 84.34% of the comments were positive, but 15.66% of consumers still expressed dissatisfaction. Follow-up comments were found to be more informative than initial comments. Among them, "positive-to-negative" comments, although accounting for a small proportion (1.02%), revealed hidden problems such as foot chafing and glue detachment during product use. In contrast, "negative-to-positive" comments highlighted the remedial value of high-quality customer service. LDA topic clustering further classified positive reviews into five themes: external quality, wearing experience, product service, product quality, and product discounts. Negative reviews focused on external quality, wearing discomfort, service dissatisfaction, quality defects, and size inaccuracy.
    Conclusion Consumers' demand for high heel shoes shows a dual characteristic of "emphasizing both appearance and comfort". Current market products have gained more recognition in terms of appearance, but there is still much room for improvement in comfort, workmanship quality, and size fitting. Follow-up reviews and the sentiment transitions within them are important sources for discovering potential product issues. Manufacturers should optimize shoe lasts based on ergonomics and real foot shape data, enhance heel structure stability, and improve the bonding process of uppers and soles. Online merchants should introduce intelligent size recommendation tools and establish active tracking and compensation mechanisms for negative customer reviews to enhance the post-purchase experience and brand loyalty. Future research can be extended to the analysis of high heel shoes reviews across multiple platforms, multi-price range, or specific functional categories.

     

/

返回文章
返回