Consumer behavior prediction and marketing strategy optimization based on big data analysis

Abstract

A large amount of user data on network platforms contains rich treasures, and the mining and development of user consumption behavior in combination with big data provide new possibilities for enterprise precision marketing. In this paper, a consumption behavior prediction model is constructed based on context-aware data and support vector machine classification algorithm, and interactive information collection is used to collect contextual data of user consumption behavior and consumption behavior cycle data. Combined with the support vector machine classification algorithm, the collected behavioral consumption data is divided into hyperplanes, and for nonlinear data, the method of placing the corresponding behavioral data in hyperplanes in the feature space is proposed. The selection of consumption behavior features is obtained using the min-max normalization process, and behavior prediction is carried out on this basis. The results show that the context-aware behavioral prediction model in this paper has the highest R-value, F1 value and NDCG among all models in Top-10, which are 0.796, 0.645 and 0.878, respectively. The two-stage prediction method in the behavioral prediction model can achieve a 98.15% data capture rate, which can accurately obtain the number of these users so as to formulate an accurate marketing strategy.

Keywords

• behavioral prediction model
• context-aware data
• support vector machine
• min-max standardization
• cross-marketing strategy
• 68p01