In the context of the deep empowerment of the digital economy in the transformation of the consumer market, consumer behavior presents digital, multidimensional, and complex characteristics. The value mining of massive consumer behavior data has become the key for enterprises to achieve precise marketing and enhance core competitiveness. Traditional consumer behavior analysis methods have limitations such as low data processing efficiency, insufficient prediction accuracy, and difficulty in capturing complex behavioral associations, which cannot meet the needs of enterprise refined operations. This article takes consumer behavior big data as the research object, integrates consumer behavior theory and big data mining technology, and constructs a complete system of consumer behavior big data processing and accurate prediction models. Firstly, we establish a multi scenario consumer behavior big data system to address the pain points of data heterogeneity and uneven quality through data collection, preprocessing, feature engineering, and dataset partitioning; Secondly, design a three-level prediction architecture of "benchmark model+integrated model+fusion optimization", selecting logistic regression and decision tree as benchmark models, XGBoost, LightGBM, CatBoost as integrated models, and combining grid search and Bayesian optimization to achieve hyperparameter optimization, constructing a model-based fusion strategy to improve prediction performance; Finally, the effectiveness of the model is verified through systematic experiments, and application strategies for the model are proposed in combination with e-commerce and retail scenarios. The experimental results show that the AUC value of the integrated prediction model reaches 0.93, and the F1 score reaches 0.91, significantly better than the single model, which can accurately predict consumer purchase behavior and loss risk. The research results of this article not only enrich the application of machine learning in consumer behavior analysis, but also provide theoretical support and practical reference for precision marketing and user management in enterprises, with important theoretical value and application prospects.

Wen Wen. Construction of Big Data Mining and Accurate Prediction Model for Consumer Behavior Based on Machine Learning. Academic Journal of Computing & Information Science (2026), Vol. 9, Issue 2: 39-45. https://doi.org/10.25236/AJCIS.2026.090206.

[1] Han H R, Jiao H F, Wang R R, Dai L Y. Progress and Prospects of the Research on Shopping Behavior of Urban Residents [J]. Progress in Geography,2011, (8):1006-1013.

[2] Zhang F, Zhang L N, Li J J. Prediction of User Consumption Behavior Based on Machine Learning Combination Model [J]. Technology of Io T& AI, 2022, (2):19-27.

[3] Qin G L. The Impact and Mechanism of Artificial Intelligence Autonomy on User Perception and Consumption Behavior—Method Based on Meta-Analysis [J]. E-Commerce Letters, 2025, (1):3806-3812.