A Study on Convolutional Neural Network Prediction Method for Consumer Purchasing Behavior in Digital Economy

This paper explores the application of deep learning techniques and their effectiveness in predicting consumer purchase behavior in the context of digital economy by constructing and evaluating convolutional neural network (CNN) models. The study is centered around a fictitious consumer behavior dataset, and describes in detail the experimental design, the model training process, the selection of performance evaluation metrics, and the comparative benchmarks with traditional machine learning methods. Through a series of fictitious experimental results, we find that the CNN model outperforms traditional models such as logistic regression, random forest, and multilayer perceptual machine (MLP) in several performance metrics, especially in terms of accuracy and F1 score. In addition, this study discusses the contribution of CNN models in terms of prediction accuracy and efficiency, as well as potential directions for future research, including model optimization, multimodal data fusion, and testing of real-world application scenarios.