The Research On Optimizing The Promotion Product Portfolio Based On User Preferences

With the development of e-commerce,online shopping has become an important part of people’s life.How to survive in the fierce market competition is a big problem.It needs companies to develop a good strategy.In the face of diversified network platform products,the traditional promotion product portfolio strategy has been difficult to applied.This paper combines with the online sales’ characteristics and presents an optimized model on the promotional product portfolio which is based on the user preferences.This model is on the basis of TOC Constraint Theory,and simplified by TOC.Then we introduced the user preference coefficient and traffic allocation model with deeply studied and analyzed.Finally,this paper verified the reliability and effectiveness of the model by an Instance simulation.Firstly,this paper elaborates the background,purpose and significance of the research,and studies the product portfolio problem,the user preference,the sentiment analysis and product recommendation which belong to the user preference.Secondly,based on the analysis of the theoretical background,this paper establishes the measurement method of user traffic model and user preference.Meanwhile it builds a promotion product portfolio model based on user preference for single activity and multiple activities.In the single activity product portfolio model,we consider the profit maximization as the target and establish the stock constraints,traffic constraints and other constraints.And then,based on the single activity product portfolio model,with the operational target of low inventory and the season factors,the product portfolio model under multiple activities is established.Besides,this paper establishes the user traffic model by principal component analysis,correlation analysis and multiple linear regression that improves the traffic utilization.And then it improves the model’s perception for the user preferences through the research of promotional factors,sentiment analysis and product recommendation which could calculate the user preference coefficient.Finally,this paper solves the promotion product portfolio under single activity and multiple activities according to particle swarm optimization algorithm and dynamic programming algorithm.It analyzes the strength and weakness with the actual situation in order to provides suggestions for enterprises’ activity planning.The promotion product portfolio based on user preference enables the promotion strategy of enterprise activities become more intelligent and convenient.In the aspect of demonstration,the model of the promotion product portfolio is superior to the actual sales of the enterprise.Especially under multiple activities,the promotion product portfolio is superior to the actual sales in the aspects of sales volume,profit and inventory.Therefore,the promotion product portfolio based on user preference not only combines with the characteristics of user preference successfully in theory,but also has great application value in practice.