Naive Bayes Classification Algorithm for Precision Marketing in E-Commerce

Precision marketing of e-commerce has become a topic of common concern for scholars at home and abroad. This paper proposes an e-commerce precision marketing model for naive Bayes and random forest combined classification algorithm. The model optimizes the naive Bayesian classification algorithm with the help of random forest algorithm. The experiment takes the data of e-commerce platform as the support, and evaluates the detection performance of the model with the help of class II confusion matrix, fitting degree and detection model skewness. The accuracy of the combination classification algorithm based on NB and RF in the four categories of beauty care, household daily necessities, mobile phones, and furniture is 85.1%, 85.4%, 72.2%, and 71.4%, respectively; The precision rates were 82.1%, 82.4%, 72.1%, and 73.4%, respectively, and the sensitivity values were 82.8%, 82.4%, 74.2%, and 74.6%, respectively. The classification model proposed in the study can accurately predict user consumption behavior in marketing, providing a scientific quantitative basis for establishing targeted marketing strategies, and also making significant contributions to the research of classification problems.