Important Considerations when Predictive Modeling with Tabular Data

My Notes for “How to Win a Data Science Competition From Top Kagglers” on Coursera.org

https://www.kaggle.com/progression

Overall, I found the course helpful and insightful, 4.79/5. There were many ideas that I had not considered before so I am posting some of my notes here. More than likely, you have seen most of these ideas so I will try to focus on the most interesting ones. Here is the link to the course.

TOC

1. Data Exploration Checklist
2. Validation
3. Target Leakage
4. Metrics and Loss Functions
5. Metric Optimization
6. Mean Encoding
7. Coding Tips
8. Advanced Feature Engineering
9. Ensemble Strategies
10. StackNet
11. Creating a Diverse Set of Models
12. Tips on Meta-Learning and Stacking
13. Text Based Features in XGBoost
14. Sequence Feature Extraction (XGBoost)
15. Semi-supervised & Pseudo Labeling
16. My Opinion

Data Exploration Checklist