Which approach usually performs best on tabular business data?
- A Gradient boosted trees such as XGBoost, LightGBM or CatBoost
- B Deep neural networks
- C K-means clustering
- D Naive Bayes
Answer
Gradient boosted trees such as XGBoost, LightGBM or CatBoost
Neural networks dominate on unstructured data such as images, audio and text, but rarely beat boosting on tabular problems.





