Titanic Survival Prediction

Project Overview

This project analyzes the Titanic dataset and predicts passenger survival using multiple Machine Learning algorithms.

Steps Performed

• Data Cleaning (handled missing values)
• Feature Engineering (FamilySize, encoding categorical variables)
• Exploratory Data Analysis
• Model Training and Evaluation

Algorithms Used

• Logistic Regression
• K-Nearest Neighbors (KNN)
• Decision Tree

Results

Algorithm	Accuracy
Logistic Regression	81.0%
KNN	81.5%
Decision Tree	78.2%

Visualizations

• Accuracy comparison bar chart
• Line graph of model performance

Key Insights

• KNN performed best on this dataset
• Feature engineering improved model accuracy
• Decision Tree showed signs of overfitting

Project Structure

• titanic_survival_prediction.ipynb → Main notebook
• train.csv, test.csv → Dataset
• images/ → Graphs and outputs

Future Improvements

• Hyperparameter tuning
• Random Forest / XGBoost models
• Deployment using Streamlit

---

Built as part of a Machine Learning project