This research conducts a comprehensive analysis of machine learning techniques used for automatic sleep stage identification. By leveraging advanced algorithms such as XGBoost, SMOTEBoost, RUSBoost, and Bagging Classifier, the study aims to optimize classification accuracy. Experimental results show that XGBoost achieved the highest classification accuracy of 83.86%, followed closely by the Bagging Classifier at 83.78%. Meanwhile, RUSBoost and SMOTEBoost recorded lower accuracies of 74.10% and 72.53%, respectively. Through extensive testing and comparative evaluation, the study highlights the effectiveness, robustness, and sensitivity of these algorithms in accurately identifying sleep stages from biomedical data. The findings offer valuable insights into selecting suitable machine-learning techniques to improve sleep monitoring, diagnosis, and healthcare. Additionally, the study underscores the importance of algorithm selection, feature engineering, and model evaluation in achieving accurate and reliable sleep stage classification.