A Nested model allows nesting of sub-models inside the main model, e.g. nesting sub-regression or sub-classification models inside a classification model.
Given a regression problem which involves molecules that can also be classified into some classes may provide the scope for first separating the molecules into those classes and then applying regression. Given such a regression problem, one may use a classifier to identify the class of an unknown sample and then apply the appropriate regression model to it. Nested Regression provides way a to combine classification and regression models. This method asks for a preliminary classification model first and then for sub-regression models.
Nested Classification provides way to combine classifier and sub-classifier models. A classification problem which involves molecules that can also be classified into some classes may provide the scope of first separating the molecules into those classes and then applying the separate class models as sub-classifier models.