A new univariate forecast method, called Xg-Boost was added to the Machine Learning Time Series forecasting ensemble. This has configurable hyper-parameters such as xgb_n_trees, xgb_depth, and xgb_learning_rate in the Zeppelin Notebook.