Enterprise Relationship Analysis Based On RFM Model And Its Application In Traceability System

With the development of information technology and product security issues,the application of product traceability system has been carried out in many countries and regions.However,the current traceability system cannot meet the needs of users to understand upstream and downstream enterprises.Meanwhile,due to numerous data,it decreases the efficiency in querying related enterprises and then influences the timeliness of product recall notifications.Therefore,we propose the enterprise relationship analysis method in the thesis,and then combine it to implement the traceability enterprise relationship analysis platform.Firstly,to realize the relationship analysis between the enterprise and its downstream enterprises,a stochastic model for valuing analysis of downstream enterprises based on RFM(Recency,Frequency,Monetary)is proposed.It utilizes the value of the downstream enterprise to evaluate the intimacy degree of the relationship between the enterprise and its downstream enterprises.The RFM model is employed to classify the downstream enterprises,and then the state-transition matrix for each kind of downstream enterprises is constructed.For each kind of downstream enterprises,the trading volume expectations of each type of downstream companies based on the average number of transactions and the expected amount of money are calculated to construct the total transaction matrix.Thereupon,the value matrix of the downstream enterprise is obtained by the product of the transfer matrix with the transaction amount matrix.The experiments illustrate that the model can effectively assess the intimacy degree between the enterprise and its downstream enterprises.Secondly,aiming at the problem that product recall notifications are inefficient,a novel approach method is proposed,which utilizes the community detection algorithm to analysis inter-enterprise relationship.In the network of enterprise relationship,the RFM model is employed to calculate the value of downstream enterprises as the weight.Then,the network is divided into different communities.It means that the same community is,the enterprises are relatively close,and vice verse.The experiments show that the results of community division can help to enhance the accuracy and efficiency of the recall notifications.Finally,combining with the method of relationship analysis between traceability enterprises,traceability enterprise relationship analysis platform is designed and implemented.It can help to perfect the function of the traceability system.Meanwhile,it can improve the comprehensiveness of traceability information and the efficiency of product recall notifications.