Research On Customer Classification Of Foreign Trade Enterprise Based On Evaluation Of Customers' Conprehensive Value

With economic globalization, enterprises are facing more and more intense competition, competition for customers become the decisive factor in the survival and development. Foreign trade enterprises is a typical customer-centric enterprise, the customer is the key factors of enterprise survival and development,80% of foreign trade enterprises' profit come from 20% of customers' orders, large customers is an important source of foreign profits, large customer marketing is also the international marketing focus. In this paper, H foreign trade company of sports protective equipment are facing how to allocate limited resources to the right customers, This paper proposed a two-step customer classification method. First, in the past, H company has accumulated a large number of customer information data, the use of support vector machine for foreign trade enterprise customers channel prediction; and build customer value and detailed channel evaluation system, using fuzzy gray relational analysis method, SVM classification results for the key channel to sort customers according to customer value. In this study, on the one hand enrich the existing customer segmentation and customer value research, while has some practical significance to the manufacturing foreign trade enterprises to analysis customer value and the rational allocation of limited resources,This paper proposed the Support Vector Machine (SVM) used in Categorized customer forecasting in foreign trade enterprise. SVM is a new pattern recognition method based on statistical learning theory,which displays many unique advantages in resolving the limited samples,nonlinear and high-dimensional pattern recognition.The main research results of this paper are as follows:Firstly, Comprehensive analyzing of the domestic and foreign customer value theory and customer classification theory based on the customer value.Then give a specific definition of the Channel distributor in this paper, research on the authority theory at home and aboard on Customer value evaluation index system,last study the history and the current situation of the SVM.Secondly, we use SVM to do Categorized customer forecasting by dividing the data in to training set and testing set, With LIBSVM toolbox completed the customer classification forecasting, finally the results classification accuracy reached 90.90%.Which proved that the rationality and feasibility of the SVM as a tool to do Categorized customer forecasting. Also we use this classification forecasting model to classify the customers of H company in the west Europe.Thirdly, to build a "multi-level performance evaluation system ", covers three areas about financial position, competitiveness, cooperation level, including 18 indicators, the paper uses the AHP Weighting Method to determine the weight because of the particularity of the foreign trade industy.Finally, combining fuzzy maths and gray system theory, this paper proposes a combination analysis method which is "Fuzzy Gray Relational Analysis" to analyze and evaluate the customer value. It uses" from low to high-level customer value analysis" t evaluate the comprehensive value of the customers. Select the six listed key channels customers of Europe to do research, using real corporate data to verify the rationality and feasibility of the model. To support the Sorting- order decision-making, giving priority to high-value customers, optimize limited resources, to achieve maximum profit.