Selection Method of Cross Border E-commerce Overseas Warehouse Partners Based on Social Networks Geolocation Prediction and ELECTRE II

Abstract This paper aims to propose a process and method to help cross border e-commerce enterprises choose overseas warehouse partners. Starting from the development tendency of cross border e-commerce, analyze the needs of cross border ecommerce enterprises for overseas warehouses. Further analyze the important demand of cross border e-commerce enterprises to choose overseas warehouse partners. Then a machine learning model that can predict user locations is obtained by utilizing publicly available data from social networks. Crawl potential consumers’ data on social networks and combine it with machine learning models to calculate the number of potential consumers in different regions. Next, based on distance and the number of potential consumers, calculate the city or city group with the lowest delivery cost. Through the actual interviews with cross border e-commerce enterprises, this paper explores the principles of cross border e-commerce enterprises’ selection and ranking of overseas warehouse partners in the same one city. In response to this decision-making problem, a model of partner selection for overseas warehouses was established based on ELECTRE II methods. The data provided by a cross border e-commerce company in Nanchang, China was used to demonstrate the selection of partners for overseas warehouses and test the effectiveness of the process. The method based on Social Networks Geolocation Prediction and ELECTRE II that this paper attempts to establish try to provide theoretical support and feasible practical solutions for the selection of overseas warehouse partners of cross border e-commerce.