Predicting repurchase behavior and optimizing marketing for e-commerce users with genetic algorithms and deep learning.

The quick expansions of e-commerce have increased competition, made client retention and repurchased predictions vital for sustainable corporate success. Traditional Deep Learning (DL) algorithms frequently fall short of collecting complex behavioral patterns and optimizing marketing campaigns. Research provides a unique framework that mixes DL and Genetic Algorithms (GA) to predict purchasing patterns and improve precision marketing for e-commerce platforms. The data of predicting e-commerce user repurchase behavior is obtained. The data features are preprocessed through handling missing values by imputation, removing duplicates, and utilizing min-max normalization for determining numerical features. A Dynamic Genetic Algorithm-mutated Enhanced Vanilla Memory Network (DGA-EVMN) is employed to enable the accurate predictions of repurchase possibilities. The EVMN model provided the temporal dependencies and nonlinear relationships in customer repurchase behavior. To further refine business decision-making, a DGA is applied to optimize personalized marketing interventions such as targeted discounts, product recommendations, and promotional campaigns. Experimental evaluation on real-world e-commerce datasets demonstrates that the proposed DGA-EVMN framework significantly outperforms traditional baseline models, achieving improvements in prediction F1-score (0.960), AUC-ROC (0.971), recall (0.962), precision (0.958) and accuracy (0.984) by the Python implementation. Moreover, the DGA-driven optimization results in more efficient resource allocation, marketing waste reduction, and higher conversion rates. The findings highlight the potential of combining predictive analytics with evolutionary optimization to enable intelligent, data-driven marketing strategies, helping businesses to maximize customer lifetime value through adaptive and cost-effective practices.