Machine learning is primarily oriented toward prediction, whereas much of economics is concerned with causality and equilibrium. While the two disciplines have a shared interest in forecasting, they often approach it with different preferences and objectives. The economics discipline tends to favor forecasting models that are explicable, parsimonious, and stable, whereas machine learning uses an empirical process for determining what is included in a model, prioritizing feature selection, regularization, and testing over intuition.
As a consequence of these ...
