Hacker News Reader

RF is a heavy hitter when it comes to tabular data. XGBoost is good as well, but more often than not needs and autotuner to really unlock it (e.g pycaret).

View on HN

XGBoost models are random forest models. They’re also just consistently better for very little effort.

you surely mean that both are ensemble models. RFs and GBMs differ in how they fit the data

A GBM like XGBoost is an ensemble of trees. It may be that when you load RandomForest modules they fit based on entropy or whatever the typical DecisionTree does but imo the term “random forest” should really convey nothing more than “ensemble of trees”.

I’m saying XGBoost would be a subclass of RF

Not only RF, they incorporate GBM too as I understand it.

Often they are the best "just run it and forget it" but compared to tuning they don't always achieve top -- sometimes surprisingly so.

XGBoost and similar are solid first stops in model building.

Reply on news.ycombinator.com