基于分类算法的移动支付系统商户采纳行为预测模型  

Prediction Model of Merchant Adoption in Mobile Payment System Based on Classification Algorithm

作　　者：李莹莹[1] 李英[1] LI Yingying;LI Ying(Business School,East China University of Science and Technology,Shanghai 200237,China)

机构地区：[1]华东理工大学商学院,上海200237

出　　处：《工业工程与管理》2021年第5期115-122,共8页Industrial Engineering and Management

基　　金：国家自然科学基金面上项目“媒体与社交网络共同影响下的信息扩散机制与动力学模型研究”(项目编号:61973121);教育部人文社科基金资助项目(15YJA630033)。

摘　　要：将Lasso-logistic回归模型与支持向量机、BP神经网络两种分类算法组合建立数据分析模型,以商户的静态社会经济属性、动态交易行为和群聚效应变量为自变量,研究商户采纳移动支付系统的影响因素,并比较了不同模型预测准确度。研究结果表明:住宿业、卫生医院业、零售业的商户更愿意采纳移动支付系统;商户的日均交易量和每笔消费金额平均数显著影响商户移动支付采纳行为;周边商户对移动支付系统的采纳占比显著正向影响商户的采纳行为。To explore the influencing factors of the adoption of mobile payment systems from the perspective of merchants,the lasso-logistic regression model was combined with two classification algorithms:support vector machine and BP neural network to establish analysis models,where static socio-economic attributes,dynamic trading behavior and clustering effect variables of merchants were used as independent variables,the prediction accuracy of different models were drawn.The results of the study indicate that:merchants in the housing industry,health hospitals and retail industries are more willing to adopt mobile payment systems;the average daily transaction volume and the average amount of each consumer significantly affect the merchant mobile payment adoption behavior;the adoption of mobile payment systems by neighboring merchants significantly positively affect the.adoption behavior of merchants.

关 键 词：移动支付 商户视角 采纳行为 机器学习 

分 类 号：F832.479[经济管理—金融学]