Machine Learning - Titanic Survival Prediction Analysis

Abstract

The most well-known shipwreck in history, the Titanic, happened in April 1912. The dataset contains comprehensive details regarding everyone that stepped onto the ship, and more than 1,500 passengers and crew members perished in the catastrophe. This study investigates the effectiveness of logistic regression, multi-layer perceptron support vector machines, and XGBoost algorithms in forecasting passenger survivability using the Titanic passenger dataset. Numerous factors, including passengers' financial status, gender, age, class, and others, influence their chances of surviving, and these factors are included in the Titanic passenger data collection. This work creates the classification model using logistic regression and other methods, preprocesses and analyzes the characteristic data, and compares the model's performance. The variables that significantly affect the survival rate are found. Experiments are used to assess each algorithm's accuracy, precision, and recall rate in order to determine which prediction model is optimal. Every algorithm has a unique optimization space and application context. In real-world use, it ought to be chosen and tailored based on the features of particular issues.