Research on Marketing and Data Analysis System Based on Computer Big Data

This paper establishes a set of enterprise marketing decision-making system based on computer big data calculation. According to the characteristics of marketing data collection and sorting, this paper proposes a mashup data architecture and a data application functional architecture. The author describes the overall functional structure of the system and the design and implementation of key processes such as the generation of marketing activities, the production of marketing activities, and the execution of marketing activities among the modules of the system. At the same time, it expounds the main functions and technical realization modes of the CRM shopping guide module, product unified management module, grid management module, marketing management module, production scheduling management module, and work division management module. This paper mines feature from the three dimensions of users, merchants, and the relationship between users and merchants in the marketing system. A total of 121 features are constructed from multiple perspectives under each dimension. In this paper, the improved Relief algorithm is used to select features with stronger discriminative ability for minority class samples and build a single model to predict the repeated purchase behavior of users. The experimental results show that after using the improved Relief algorithm to select features, the predicted effect of the model is better.