We analyzed the efficiency of four machine learning models—ANN, LSTM, Decision Tree, and Linear Regression—in demand forecasting using a textile company's sales data. ANN provided the most accurate forecasts with the lowest data loss, while LSTM was slightly less accurate. These findings highlight the importance of advanced models in optimizing inventory management and supply chain efficiency.

This study compares Temporal-DBN, Variational-DBN, and Sparse-DBN models for anomaly detection in business intelligence, finding Temporal-DBN significantly outperforms with a 97.67% accuracy, indicating its potential to enhance data-driven decision-making in diverse business fields.