Research of E-commerce Resource Allocation Method Based on Clustering

This paper focuses on the research of diversified e-commerce activities in e-commerce platform, mapping various business activities into the framework of resource allocation. When a certain number of users access the platform concurrently, the system allocates resources. This paper abandons the traditional simple filtering mode to analyze the historical data of e-commerce platform, abstracts the user commodity binary network and divides the user communities according to the attribute information of users and commodities and the purchase logs of users. By analyzing the community of users, we can allocate resources reasonably. In the binary network of users and goods, there are still many problems in user based clustering. For example, the sparseness of user attribute information, the proportion of intersection between users' purchasing goods is too small and so on. According to the characteristics of data and the existing problems, this paper selects the SimRank algorithm and improves it. In this paper, a combined similarity index based on user attributes and network structure is proposed to improve the accuracy of user clustering, and a reasonable scheme of resource allocation is proposed by using the community division of users, which increases the accuracy of user clustering, achieves the expected design goals, and improves work efficiency.