The Titanic project

Introduction

Kaggle's Titanic: Machine Learning from Disaster is one of the most popular beginner-friendly projects for learning data science and machine learning. The goal of the project is to predict whether a passenger survived the Titanic shipwreck based on their characteristics, such as age, gender, ticket class, and more.

The raw pages from Kaggle are here.

Table of Contents

• .gitignore
• gender_submission.csv
• submission.csv
• test.csv
• train.csv
• test_data.json
• train_data.json
• process_data.py
• Training.py
• Trainingprocess.ipynb

Explanations

The file gender_submission.csv, test.csv and train.csvare downloaded from the Kaggle website.

The file submission.csv is the final output result of the prediction, which has been committed to the Kaggle.

The process_data.py is used for transforming the csv files to json files, including test_data.json and train_data.json.

The Trainingprocess.ipynb is used for training the json data and getting the output.

The Training.py has the same content with Trainingprocess.ipynb.

Implementation

The training process begins with data cleaning and preprocessing, which involves converting strings and filling in missing data. Next, feature extraction is performed on certain variables. After that, the Random Forest algorithm is used to train the model, and the prediction accuracy is evaluated while a confusion matrix is plotted. Finally, the trained model is applied to the test set to output the final results.

Usage

The project can be an excellent experience to embark on the first machine learning project.

Please ensure you have installed several packages below:

pip install pandas
pip install numpy
pip install sklearn

After the installation, you can train the model using commands below:

git clone https://github.com/xiyuanyang-code/Titanic.git
cd Titanic
python process_data.py
python Training.py

Moreover, I suggest to run the Trainingprocess.ipynb in the Visual Studio Code, where you can run every cell independently to implement each function accordingly.

Advertisement

My personal Blog