UCLA

The factors influencing survival rates during disasters had always been an important subject of research. With the rise of machine learning, predictive modeling has improved significantly. This paper presented a comparative analysis of three Machine Learning models—XGBoost, Random Forest, and AdaBoost—trained using well-established libraries to predict the survival probabilities of passengers on the Titanic. We used a well- known dataset from the Titanic disaster, containing passenger information and whether they survived. After data preprocessing and model tuning, Random Forest showed the highest accuracy, suggesting its potential for improving survival predictions in disaster rescue operations.