Which approach is generally strongest for tabular business data?
- A Gradient boosting implementations such as XGBoost, LightGBM or CatBoost
- B Deep convolutional networks
- C Recurrent neural networks
- D K-means clustering
Answer
Gradient boosting implementations such as XGBoost, LightGBM or CatBoost
Well-tuned gradient boosting usually outperforms neural networks on tabular problems, which is a useful practical point to make in interviews.





