Poster presented in the NSF-AGEP Research Exchange Retreat (2019) at Stanford University.
The poster pdf can be found here: Poster.
Applicability Domain in Data Science
There are times when a model's prediction should be taken with some skepticism. For example, if a new data point is substantially different from the training set, its predicted value may be suspect. In chemistry, it is not uncommon to create an "applicability domain" model that measures the amount of potential extrapolation new samples have from the training set. The methods used to define the applicability domain of a model can be applied to data sets not necessarily derived from chemistry. Here we describe various methods used to define the applicability domain of any model. In addition, we present the modeling R package
applicable, which comprises different methods to measure how much a new data point is an extrapolation from the original data.