Intelligent Construction of E-commerce Product Traceability System Based on Big Data Mining

With the development of network technology and information technology, the Internet has inexorably entered people's life, and changed people's way of life. However, due to the virtual and trans-regional nature of e-commerce, the products purchased by consumers may have quality risks. As an effective supervisory tool, traceability is an effective measure to ensure the quality of e-commerce products. At present, the development of traceability in foreign countries is relatively mature, and specific industries such as food and agricultural products are also widely applied in China. However, at home and abroad, the application of traceability in e-commerce is still relatively small. Therefore, in view of this problem, this paper aims to explore the correlation between e-commerce product data through correlation analysis, from the based on the correlation analysis and trend prediction of the base data, the potential rules and patterns are found through analysis, and finally the network model is constructed to meet the actual needs of risk management, and the traceability efficiency evaluation and traceability algorithm design are realized. The experiment proves that the improved Bayesian algorithm is superior to the NB algorithm. After feature selection and dimension reduction, the model training of the improved Bayesian algorithm is carried out, and its accuracy reaches the highest at 18,500 dimensions. Applying the improved weighted Bayes algorithm to the quality risk control assessment of e-commerce goods can better predict the quality risk level of e-commerce goods.