Study On The Method For Customer Behavior Prediction Based On Cox Model Under Mobile Commerce

The rapid development of mobile communication technology,makes people's communication and network life get rid of the bondage of the geographical and time,and be able to access to the network at anytime and anywhere.The revolutionary changes also led to the electronic commerce activities extend from wired network to wireless network,and greatly promoted the emergence and development of mobile commerce.There are two different characteristics between mobile commerce and traditional electronic commerce:position sensitivity and time sensitivity.Because of the time sensitivity,mobile commerce enterprises must timely grasp the accurate time point and place that customers' trading activity or the service request may happen at in the future to ensure to maximize customer value.How to consider both exploration of customer behavior rules and time sensitivity of mobile commerce,is a problem to be solved.Therefore,study on the method for customer behavior prediction under mobile commerce is a realistic research topic.Current researches about this problem has caused extensive concern of the business community and academia,and some research achievements have been made.However,the existing researches have deficiencies both in depth and breadth aspect.First of all,on the depth of the study,the attention to time sensitivity of mobile commerce is insufficient and can't meet the needs of enterprise that grasp the time point of customers' trading activity happened at.Secondly,on the breadth of the study,the existing research in view of the existing customers with historical records,and less relating to the potential customer behavior prediction problem.Therefore,study on the method for customer behavior prediction under Mobile Commerce is needed.Fully considering the time sensitivity,giving a method for existing customer and potential customer behavior prediction to predict the happened time and results of customer behavior.The informed research has shown that Cox model has certain advantages in terms of customer behavior prediction,analysis and time early warning capacity.Therefore,it's very necessary to introduce the method for customer behavior prediction based on Cox model into customer behavior prediction problem under mobile commerce.In this paper,we carried on a thorough study on the method for customer behavior prediction based on Cox model under mobile commerce.A series of research works are conducted as follows.(1)The problem of customer behavior prediction based on mobile commerce is described and the theoretical framework for the research problem is proposed.Mobile commerce customers can be divided into existing customers and potential customers according to whether has relevant historical behaviors in the product category.Then customer behavior prediction problem under mobile commerce is divided into existing customer behavior prediction and potential customer behavior prediction problems.In addition,the research problem is described,and the research framework for solving the problem is proposed.These basic research works mentioned above provide a theoretical guidance framework and analytical frame work to the related problems and lay a solid foundation to extended issues and applications.(2)The method for existing customer behavior prediction based on Cox model is proposed.First of all,based on the historical records of existing customers in product or product category,established Cox model to predict the result of the customer behavior occurs.Then,based on analysis results of Cox model,established a method for time point prediction to predict the time point of the existing customer behavior occurs;Finally,proposed an evaluation method to analyze the prediction effect.(3)The method for potential customer behavior prediction based on Cox model is proposed.First of all,based on the historical records of potential customers in related product or product category,established Cox model to predict the result of potential customer's first behavior occurs.Then,established a method for time point prediction to predict the time point of the potential customer behavior occurs based on analysis results of Cox model.Finally,proposed an evaluation method to analyze the prediction effect.The proposed method for customer behavior prediction based on Cox model considers the time sensitivity of mobile commerce.It provides a reference method to customer behavior prediction problem under mobile commerce.