This topic explains how the NGBoost model works


The NGBoost model provides Stanford ML Group's NGBoost. NGBoost enables predictive uncertainty estimation with Gradient Boosting through probabilistic predictions (including real valued outputs). With the use of Natural Gradients, NGBoost overcomes technical challenges that make generic probabilistic prediction hard with gradient boosting.

Probability Distributions

Myst's implementation of NGBoost contains four probability distributions - Normal, Cauchy, Laplace, and Lognormal. Each model outputs scale and loc parameters that can be used to determine the upper and lower bounds of the associated probability distribution. The loc parameter in the lognormal is returned in log space.

Backtest Results

In backtests MAPE, MSE, and MAE are reported for symmetric distributions (Normal, Cauchy, Laplace). For asymmetric distributions (Lognormal), only mean-negative-log-likelihood.


ParameterDescriptionDefault Value
distributionThe assumed distributional form of the endogenous data, given the exogenous data.Normal
n_estimatorsThe number of boosting iterations to fit.500
learning_rateThe learning rate0.01
max_depthThe maximum depth of the base learner.3
minibatch_fracThe percent subsample of rows to use in each boosting iteration.1
col_sampleThe percent subsample of columns to use in each boosting iteration.1


See our Create a Price Forecast with Probabilistic Prediction tutorial for examples using NGBoost!