Predict which passengers are transported to an alternate dimension
This project contains my entry for the spaceship-titanic Kaggle competition.
It currently gets a 79% score. I'll take it :).
ID | Score | Description | Leaderboard position |
---|---|---|---|
0 | 78% | Balanced | |
1 | 77% | Unbalanced | |
2 | 79% | Balanced, with cabin data | 1485 |
- Explain what parameters matter!
- Smarter data filling
- Better feature engineering
- Exploit Cabin data
- Exploit first name and last name
- Hyperparameter tuning for the model
- Testing with and without balancing -> balanced model performs 0.1% better compared to unbalanced