🚢 Titanic Survival Prediction

A machine learning project that predicts passenger survival on the Titanic using various classification algorithms with hyperparameter optimization.

📋 Description

This repository contains a comprehensive machine learning solution for the famous Kaggle Titanic survival prediction competition. The project implements multiple classification algorithms with hyperparameter optimization to predict whether a passenger survived the Titanic disaster based on features like age, gender, ticket class, fare, cabin, and more.

The solution employs a systematic approach:

1. Data exploration and visualization
2. Feature engineering and preprocessing
3. Model training with hyperparameter optimization
4. Model evaluation and selection
5. Prediction generation for test data

✨ Features

• Exploratory Data Analysis: Comprehensive analysis of the Titanic dataset with visualizations to understand feature relationships and survival patterns
• Feature Engineering: Creation of new features like family size, title extraction from names, and family survival correlation
• Hyperparameter Optimization: Uses Hyperopt library to find optimal parameters for each model
• Multiple Classification Algorithms:
  • Decision Tree Classifier
  • Random Forest Classifier
  • Gradient Boosting Classifier
  • XGBoost Classifier
  • K-Nearest Neighbors
  • Support Vector Machine (implementation available but not used in main script)
  • Neural Network with Keras (implementation available but not used in main script)
• Model Comparison: Automatic selection of the best performing model for final predictions

🛠️ Setup

Prerequisites

• Python 3.x
• Required libraries:
  • pandas
  • numpy
  • scikit-learn
  • matplotlib
  • seaborn
  • hyperopt
  • xgboost
  • keras (optional, for neural network implementation)

Installation

1. Clone this repository:

git clone https://github.com/yourusername/KaggleTitanticSurvivalClassify.git
cd KaggleTitanticSurvivalClassify

2. Install required packages:

pip install pandas numpy scikit-learn matplotlib seaborn hyperopt xgboost keras

📊 Usage

1. Run the main script to train models and generate predictions:

python titanicPredictSurvival.py

2. The script will:
   • Load and preprocess the training and test data
   • Perform feature engineering
   • Train multiple models with hyperparameter optimization
   • Select the best performing model
   • Generate predictions for the test set
   • Save predictions to test_set_prediction.csv

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.