Prediction of the Titanic Survival Probability Based on Boosting Strategies

Abstract

Listen

Translate article

According to the unsatisfactory rescue outcomes in natural disasters such as marine accidents and the challenges in survival prediction at nowadays, this study employs Gradient Boosting, XGBoost model, and Optimized XGBoost model (based on RandomizedSearchCV method) to forecast the survival of Titanic passengers. The research begins by processing data on various passenger characteristics from the Titanic. Through comparative analysis of experimental results from these three predictive models, it manifests that the Optimized XGBoost model demonstrates highest precision and accuracy, while being less prone to overfitting and underfitting.