Research on New Media Precision Marketing Method Based on Big Data Information Automatic Push

Abstract

Aiming at the problems of low precision and overload of commodity push system, this paper proposes a precise marketing push model based on user portrait feature information. Based on the analysis results and scoring matrix of users' historical behaviors, the user's interest model is constructed, and it is transformed into the user's label model by using label rules. At the same time, the Thrift framework interface is called according to a variety of mixed rules, and the user's preferred push commodity list is returned, so that an accurate marketing push system with excellent performance and low delay time is realized. The simulation results show that the precision marketing push system in this paper uses NDCG algorithm to optimize the parameters, which has high accuracy and stability.