Choosing Between Gradient Boosting and Neural Networks for Tabular Data
Deep learning has transformed computer vision, natural language processing, and speech recognition. Yet for structured, tabular data such as spreadsheets, transaction records, and relational database exports that represent much of real-world business data, gradient boosted decision trees (GBDTs) such as XGBoost, LightGBM, and CatBoost often outperform neural networks. Understanding why...
