This repository contains my Kaggle competition notebooks and solutions.
- Rank: 182 out of ~6,000 submissions
- Score: 0.11838
- Approach: Ensemble learning combining Random Forest, Gradient Boosting, Elastic Net, and Kernel Ridge models
- Competition Link
- Score: 0.78947
- Approach: Random Forest with extensive feature engineering
- Features:
- Title extraction from names
- Family size calculations
- Age and fare binning
- Cabin deck extraction
- Ticket prefix analysis
- Interaction features
- Competition Link
- Approach: Convolutional Neural Network (CNN)
- Features:
- 3-layer CNN architecture
- Hyperparameter optimization
- Data normalization and reshaping
- Comprehensive visualization
- Competition Link
- Extensive data preprocessing and feature engineering
- Model ensembling and stacking
- Hyperparameter optimization
- Cross-validation techniques
- Python
- scikit-learn
- pandas
- numpy
- matplotlib
- seaborn