Titanic Disaster Analysis

The Titanic Problem

Here we have a Titanic Problem which is part of the Kaggle competition. This project involved predicting the number of survivors in the Titanic disaster using machine learning by building an accurate model. It has various features like Passenger ID, Age, Sex, the cabin numbers of the passengers,etc. The model we have used here is Random Forrest Tree which works on the principle of having most votes. After running the tree on our dataset, we get an accuracy of about 0.77511.

Contribution

The accuracy obtained was about 0.77511 so in order to increase it certain changes and steps had to be taken. Firstly taking care of NA values. This was done by replacing all NA values from the 'Age' column with the mean of the 'Age' column. After this, in the Random Forrest model, more features were added and some changes were also made in the depth and estimators of the tree. By performing these changes, an accuracy of about 0.78 was obtained.

TitanicProb (2).ipynb

Download IPYNB • 8KB