Boosting Techniques for Time Series Forecasting

We're going to talk about boosted and bettings. But maybe to set the stage for listeners who have a passing familiarity, could you give a high level definition of what boosting is? So there we have many weak learners that can estimate some function with some propablty rictlet say, half plesepsilon. And now you want to put them all together. The way you do it in a more gentle framework is like you keep successively fitting these weak learners and seeing on which inputs are they doing bad.