Titanic-ML

Titanic Survival Prediction

Project Overview

This project presents a complete machine learning workflow to predict passenger survival on the Titanic. . The goal was to go beyond a basic model and demonstrate a data-driven approach, from initial exploratory analysis and data cleaning to a comparative evaluation of multiple classification algorithms to identify the most effective model.

Methodology

The project was broken down into three key phases:

1. Data Visualization & Analysis:

   • Performed a comprehensive Exploratory Data Analysis (EDA) to understand the dataset's structure and uncover key relationships between passenger attributes (like class, sex, and age) and survival rates.
   • Created several visualizations, including correlation heatmaps and bar charts, to inform feature selection and preprocessing strategies.

2. Data Cleaning & Preprocessing:

   • Executed robust data cleaning procedures, handling missing values in Age, Fare, and Embarked with mean/mode imputation to ensure a complete dataset for modeling.
   • Transformed categorical features like Sex and Embarked into numerical formats to make them compatible with machine learning algorithms.

3. Model Training & Comparison:

   • Trained and evaluated three distinct classification models to find the best predictor for this task:
     • K-Nearest Neighbors (KNN)
     • Decision Tree Classifier
     • Random Forest Classifier
   • Rigorously tested each model on unseen data to ensure the results were generalizable and reliable.

Results

The Random Forest Classifier delivered the highest accuracy, demonstrating the strength of ensemble methods on this dataset.

Model	Accuracy
K-Nearest Neighbors	72.22%
Decision Tree Classifier	77.78%
Random Forest Classifier	81.11%

Technologies Used

• Python
• Pandas
• NumPy
• Matplotlib
• Seaborn
• Scikit-learn

Dataset

• Source: Kaggle - Titanic: Machine Learning from Disaster
• File used: Titanic-Dataset.csv

Titanic Model Comparision

• Three Classifiers are compared - KNearesKNeighborsClassifier, DecisionTreeClassifier and RandomForestClassifier

Titanic Data Visualization

• Various models like - Barplot, heatmap, and pairplot are used to understand and visualize the data.

How to Run

1. Clone the repository.
2. Ensure you have the required libraries installed (pandas, matplotlib, seaborn, scikit-learn).
3. Open and run the Jupyter Notebooks (Titanic.ipynb, TitanicModelComparision.ipynb, TitanicVisualization.ipynb) to see the full analysis and model training process.

---

Dataset License Notice

The Titanic dataset used in this project is publicly available for educational use and is provided by platforms like Kaggle.

---

Created by Deepak Battula