The Titanic accident is a remarkable historical loss in the cruise industry with only 32% survival rate. In this project, a binary classification model is built to predict whether a Titanic accident passenger survived or not using python programming language. The paper analyzes 11 variables of each customer to find out which factors have the most significant effect on the customers’ survival rate and then uses binary classifiers for prediction. The process involves identifying relevant features, highlighting and visualizing the interesting relationships between them, and that with the target variable (survived). We predict survival using Random Forest, Logistic Regression, Naive Bayes, k-nearest neighbors(KNN) and Decision Trees. We evaluate the performance by using a confusion matrix to calculate recall, accuracy, f1-score and precision. The results from the paper will help cruise companies have better planning to protect and rescue people from similar accidents. It will also help customers to make an appropriate decision to increase their chances of survival in case of a similar incident.
Keywords: Binary Classification, Survived, Predictive Models, Algorithms, Logistic Regression, Decission Trees, Random Forests.