Redeeming Data Science by Decision Modelling

TL;DR

This paper advocates for a new approach called Decision Modelling that integrates causal graphical models and value frameworks to ground Data Science practices in AI principles, emphasizing decision quality.

Contribution

It introduces Decision Modelling as a novel framework combining ML models with explicit value models, grounded in Bayesian and AI techniques, to improve Data Science applications.

Findings

Decision Modelling integrates ML with value models like utility.

It applies causal graphical models to Data Science.

The framework emphasizes six principles of Decision Quality.

Abstract

With the explosion of applications of Data Science, the field is has come loose from its foundations. This article argues for a new program of applied research in areas familiar to researchers in Bayesian methods in AI that are needed to ground the practice of Data Science by borrowing from AI techniques for model formulation that we term ``Decision Modelling.'' This article briefly reviews the formulation process as building a causal graphical model, then discusses the process in terms of six principles that comprise \emph{Decision Quality}, a framework from the popular business literature. We claim that any successful applied ML modelling effort must include these six principles. We explain how Decision Modelling combines a conventional machine learning model with an explicit value model. To give a specific example we show how this is done by integrating a model's ROC curve with a utility model.