Ethical Considerations In Parsimonious Model Selection
Model algorithms are frequently being used behind the scenes to make decisions on our behalf. We should all care about their pervasive use and impact on society. Consequently, we all need to be savvy in understanding the fair use of models and their parsimonious selection.
Ethical Considerations in Parsimonious Model Selection, the newest PRMIA Institute paper written by Dr. Gary Nan Tie and Dr. Bob Mark, explores how fairness is an important consideration in finding a parsimonious model solution. Good practice calls for first deciding what you want to accomplish in constructing a model. Accordingly, it is important to first write down the question you are trying to answer. In articulating purpose, we need to consider ethical issues. You next find a parsimonious solution given that you now know what you want to accomplish.
A parsimonious model choice embodies an economy of conceptualization wherein we avoid either being unnecessarily elaborate or too simplistic. We choose a parsimonious model in terms of only that which is needed to understand our problem and to robustly extrapolate. By doing so, we find solutions to better understand and mitigate model risk.