In today's digital age, with an increasing number of online interactions and behavioral data generated by consumers, predicting consumer purchase intent has become a vital part of personalized digital marketing. This study offers an AI-based personalized digital marketing framework that incorporates data normalization and NLP-based text cleaning methods to boost data quality, and derive actionable insights from customer reviews, browsing behaviors, and transaction logs. To solve the computational problem, SHAP-based feature importance selection is used to select the most important factors that influence customers' purchase decisions. The developed Graph Neural Network combines a Transformer-based classifier to obtain an accurate purchase intention prediction by capturing complex consumer-product relationships and long-term consumer behavioral patterns. The framework is realized with Python and PyTorch Geometric for scalable graph learning and deep representation modeling. It is realized in terms of a framework written in Python and PyTorch Geometric for scalable graph learning and deep representation modeling. The experimental results show promising predictive capabilities, with high accuracy, precision, recall and ROC-AUC scores over the traditional machine learning and deep learning methods. The proposed system can enhance personalization, customer interaction, and marketing effectiveness, which is applicable for intelligent real-time digital marketing environments.
Download
Download XML