Optimizing Employee Engagement and Retention through Personalized E-Commerce Solutions with Predictive Analytics Using CNN-Autoencoder for HR Decision Making

Employee retention and engagement are key drivers of organizational performance. However, traditional approaches frequently miss the complexities of contemporary work patterns. This paper outlines a framework that uses customized e-commerce solutions powered by predictive analytics to improve employee retention and engagement. The framework employs a hybrid model combining Convolutional Neural Networks (CNN) and Autoencoders. It makes predictions from comprehensive datasets that include employee behavior, performance, and engagement levels. Through a hybrid two-stage feature extraction pipeline combining PCA-based linear dimensionality reduction and CNN-Autoencoder encoder-based nonlinear feature extraction, the model identifies engagement levels and predicts disengagement. The model performance was assessed using several key metrics. The results are: Accuracy = 99.48%, Precision = 99.34%, Recall = 99.63%, F1-Score = 99.49%, MAE = 0.0052, RMSE = 0.0719, and R 2 = 0.9793. These findings highlight the model's strong predictive capability for HR decision-making. It enables targeted, data-driven interventions that improve employee satisfaction, reduce turnover, and boost productivity. The model shifts HR practices from reactive to proactive. This helps organizations retain top talent and maintain a high-performing, motivated workforce in a competitive labor market.