The notebook includes the following:
- Data Overview
- Exploratory Data Analysis and Data Cleaning
- Data Visualizations
- Handling Missing Values
- Feature Engineering
- Encoding
- Managing Multicollinearity and Correlation Analysis
- Model Training and Selection
- Final Model
Nested Cross-Validation was performed for the following models:
- K-Nearest Neighbors
- Random Forest
- Support Vector Machine
- AdaBoost
- Gradient Boosting
- XGBoost
After extensive tuning, the final model achieved an out-of-bag (OOB) accuracy score of ≈0.81 and an accuracy of 0.80382 on the private test dataset, placing it in the top 3% among competition participants.
| Notebook | Leaderboard Score | Rank |
|---|---|---|
| titanic notebook | 0.80382 | Top 3% |