摘要
【背景和目的】针对传统机器学习方法在中高端卷烟消费者多动机识别中的局限性,本文通过构建基于支持向量机(Support Vector Machine,SVM)的瀑布模型分析方法,以提升对消费者多动机的识别准确率,进而高效指导烟草企业制定精准营销策略。【方法】基于用户行为对营销策略的导向,消费者行为动机分为交往型、追新型、求异型、从众型、务实型5类。依据前述5类行为动机在中高端卷烟营销实践中的优先级不同,本文构建三级SVM瀑布模型分类器。【结果】实验表明,所提出模型的综合准确率达92.67%,较传统方法提升约10%或以上。【结论】在安徽省合肥市使用该模型筛选出130家店铺,并对其进行中高端卷烟营销,结果显示,这些店铺中高端卷烟订购总量上升了29.8%,表明模型在实际营销中具有显著价值,推广前景广阔。
[Objective]To address the limitations of traditional machine learning methods in recognizing multiple behavioral motivations of premium cigarette consumers,this study proposes an SVM-based cascade model to improve recognition accuracy and provide effective guidance for tobacco enterprises in developing precise marketing strategies.[Methods]Based on the orientation of user behavior toward marketing strategies,consumer behavioral motivations were classified into five categories:communicative,novelty-seeking,divergent,conformity,and pragmatic.According to the different priorities of these five motivations in the marketing practice of premium cigarettes,a three-level SVM cascade classifier was constructed.[Results]Experimental results demonstrated that the proposed model achieved an overall accuracy of 92.67%,representing an improvement of more than 10%compared with traditional methods.[Conclusion]Applying the model to screen 130 retail stores in Hefei,Anhui Province,and implementing premium cigarette marketing in these stores,the total order volume of premium cigarettes increased by 29.8%.These findings indicate that the model provides significant practical value for marketing applications and has broad prospects for wider promotion.
作者
蒋晨辰
孔维力
黄丽莹
张子超
韦伟
李纲彬
张磊
宋文慧
JIANG Chenchen;KONG Weili*;HUANG Liying;ZHANG Zichao;WEI Wei;LI Gangbin;ZHANG Lei;SONG Weihui(China Tobacco Guangxi Industrial Co.,Ltd.,Nanning 530000,China;Shenzhen Wentong Technology Co.,Ltd.,Shenzhen 518000,China)
出处
《中国烟草学报》
北大核心
2025年第6期143-151,共9页
Acta Tabacaria Sinica
基金
广西中烟工业有限责任公司科技项目“安徽市场中高端卷烟消费者行为研究”(CGAXZX20220030050001-049)。
