医药电商行业消费者满意度关键影响因素识别---基于抗感冒药品在线评论的研究

Identifying factors influencing consumer satisfaction in the pharmaceutical e-commerce sector: A study based on online reviews of Anti-Cold Drugs

摘要

本文以消费者满意度为切入点,聚焦医药电商平台服务质量优化路径,为精准提升服务质量提供数据支撑.基于60396条抗感冒药品有效评论数据,构建文本挖掘、主题特征提取、影响因素排序三维分析框架.运用LDA主题模型确定最佳主题数量,并挖掘评论文本潜在主题,识别出物流配送服务、药品价格等7个核心影响因素.进而以评论的文档主题概率分布向量为输入特征,构建XGBoost分类模型,通过与逻辑回归、随机森林、支持向量机、CatBoost等模型在准确率等5类性能指标上的对比,证实该模型对数据具有显著适配性,经参数优化后,模型性能达到最优水平,并输出特征的重要性得分.研究结果表明不同剂型药品的消费者满意度影响因素存在显著差异,从而为医药电商消费者行为研究提供有效的理论支撑.

Abstract

This paper takes consumer satisfaction as the entry point and focuses on the optimization path of service quality in pharmaceutical e-commerce platforms, aiming to provide data-driven support for the precise improvement of service quality. Based on 60396 valid review entries for anti-cold medications, a three-dimensional analytical framework encompassing text mining, topic feature extraction, and influencing factor ranking was constructed. The LDA topic model was employed to determine the optimal number of topics and to uncover latent themes within the review texts, identifying seven core influencing factors, including logistics delivery service and drug pricing. Subsequently, using the document–topic probability distribution vectors from the reviews as input features, an XGBoost classification model was developed. By comparing its performance against models such as logistic regression, random forest, support vector machine, and CatBoost across five performance metrics (e.g., accuracy), the proposed model was demonstrated to exhibit superior adaptability to the dataset. After parameter optimization, the model achieved optimal performance, and the importance scores of the features were output. The findings reveal significant differences in the factors influencing consumer satisfaction across different dosage forms of medications, thereby providing robust theoretical support for research on consumer behavior in pharmaceutical e-commerce.

关键词

消费者满意度 / 在线评论 / LDA主题模型 / XGBoost / 影响因素

Key words

consumer satisfaction / online reviews / LDA topic model / XGBoost / factors

参考文献

[1] Oliver R L . A cognitive model of the antecedents and consequences of satisfaction decisions[J]. Journal of Marketing Research, 1980, 17(4): 460-469.

[2] Churchill G A Jr, Surprenant C . An investigation into the determinants of customer satisfaction[J]. Journal of Marketing Research, 1982, 19(4): 491-504.

[3] 王宁. 数字化背景下乡村旅游消费者满意度测评模型构建[J]. 商业经济研究, 2021(8): 57-60.

[4] Wang N. Construction of rural tourism consumer satisfaction evaluation model under the background of digitalization[J]. Journal of Commercial Economics, 2021(8): 57-60.

[5] 廖名岩, 王璜, 张世兵, 等 . 消费者购买可数字化追溯猪肉的影响因素[J]. 系统工程, 2025, 43(4): 27-38.

[6] Liao M Y, Wang H, Zhang S B, et al. The factors influencing consumer purchase of digitally traceable pork[J]. Systems Engineering, 2025, 43(4): 27-38.

[7] 朱郭奇, 钱冬, 李文文, 等 . 社交电商平台消费者忠诚建立:对信任与满意度的再认识[J]. 管理工程学报, 2025, 39(5): 74-91.

[8] Zhu G Q, Qian D, Li W W, et al. Building consumer loyalty to social e-commerce platforms: a re-understanding of trust and satisfaction[J]. Journal of Industrial Engineering and Engineering Management, 2025, 39(5): 74-91.

[9] 陈丁丁, 崔荣荣, 牛犁. 基于Kano模型的裘皮服饰设计需求研究[J]. 丝绸, 2025, 62(12): 16-25.

[10] Chen D D, Cui R R, Niu L. Research on fur clothing design requirements based on the Kano model[J]. Silk, 2025, 62(12): 16-25.

[11] 李敏, 姜夏楠, 陈彩霞. 服装品牌智能客服服务满意度的影响因素研究[J]. 丝绸, 2025, 62(9): 33-42.

[12] Li M, Jiang X N, Chen C X. Research on the influencing factors of artificial intelligence customer service satisfaction of fashion brands[J]. Journal of Silk, 2025, 62(9): 33-42.

[13] 熊百妹. 基于逐步回归法的药品消费者满意度影响因素实证研究[J]. 中国药房, 2013, 24(17): 1548-1550.

[14] Xiong B M. Empirical study on influential factors to drug customer satisfaction based on stepwise regression method[J]. China Pharmacy, 2013, 24(17): 1548-1550.

[15] Zhao X Q, Huang Z . A method for exploring consumer satisfaction factors using online reviews: a study on anti-cold drugs[J]. Journal of Retailing and Consumer Services, 2024, 81: 103895.

[16] 倪昭鑫, 舒帆. 基于自注意力机制的CNN-BiLSTM生鲜物流服务质量影响因素[J]. 科学技术与工程, 2025, 25(16): 6821-6830.

[17] Ni Z X, Shu F. Factors affecting fresh logistics service quality based on self-attention mechanism of CNN-BiLSTM[J]. Science Technology and Engineering, 2025, 25(16): 6821-6830.

[18] 邬先利, 廖虎昌. 多平台在线评论主题挖掘与消费者偏好演化趋势分析[J/OL].[ 2025-07-10]. https://link.cnki.net/urlid/31.1977.N.20250606.1437.002.

[19] Wu X L, Liao H C. Topic mining and consumer preference evolution trend analysis based on online reviews from multiple platforms[EB/OL].[ 2025-07-10]. http://en.cnki.com.cn/Article_en/CJFDTOTAL-XTGL20250605001.htm.

[20] 何勇, 陈静, 李姗姗, 等 . 生鲜电商用户在线选择行为研究:考虑评论的影响效用[J]. 管理科学学报, 2025, 28(5): 70-83.

[21] He Y, Chen J, Li S S, et al. Online choice behavior for fresh e-commerce customers: Considering the social influence of reviews[J]. Journal of Management Sciences in China, 2025, 28(5): 70-83.

[22] Krishna B, Puram P . The impact of trust-based challenges on user satisfaction in food sharing platforms: a text mining approach[J]. Technological Forecasting and Social Change, 2025, 216: 124159.

[23] 李伟, 徐华, 李媛媛. 节假日时段国内自驾游客满意度研究:基于结构方程模型的实证分析[J]. 河南师范大学学报(自然科学版), 2020, 48(6): 1-9.

[24] Li W, Xu H, Li Y Y. Satisfaction of domestic self-driving tourists during holidays: Empirical analysis based on SEM[J]. Journal of Henan Normal University (Natural Science Edition), 2020, 48(6): 1-9.

[25] Wang G Q, Zhu B W, Wang K Y, et al. Links between cue combinations of physical environments and consumer satisfaction in themed restaurants from a systematic approach-avoidance perspective[J]. International Journal of Hospitality Management, 2025, 126: 104055.

[26] 吴东雪, 沈桂兰. 一种基于LDA模型的新兴主题识别与探测方法[J]. 河南师范大学学报(自然科学版), 2024, 52(2): 72-80.

[27] Wu D X, Shen G L. An emerging topic identification and detection method based on LDA model[J]. Journal of Henan Normal University (Natural Science Edition), 2024, 52(2): 72-80.

[28] 蒋昊达, 赵春蕾, 陈瀚, 等 . 基于改进TF-IDF与BERT的领域情感词典构建方法[J]. 计算机科学, 2024, 51(S1): 150-158.

[29] Jiang H D, Zhao C L, Chen H, et al. Construction method of domain sentiment lexicon based on improved TF-IDF and BERT[J]. Computer Science, 2024, 51(S1): 150-158.

[30] 韩坤, 刘忠轶. 基于主题模型与情感特征增强的食品安全网络舆情演化分析方法[J]. 食品科学, 2025, 46(23): 367-375.

[31] Han K, Liu Z Y. A method for analyzing the evolution of online public opinion on food safety based on topic modeling and sentiment feature enhancement[J]. Food Science, 2025, 46(23): 367-375.

[32] Chen T Q, Guestrin C . XGBoost: A Scalable Tree Boosting System[C]. New York: Association for Computing Machinery, 2016: 785-794.

[33] 温廷新, 白云鹤. 融合RF-GA-XGBoost和SHAP的虚假新闻群体互动质量可解释模型[J]. 数据分析与知识发现, 2024, 8(12): 86-100.

[34] Wen T X, Bai Y H. An interpretable model for fake news Groups’ Interaction quality based on RF-GA-XGBoost and SHAP[J]. Data Analysis and Knowledge Discovery, 2024, 8(12): 86-100.

基金资助

国家自然科学基金(62176050)