Considerations Across Three Cultures: Parametric Regressions, Interpretable Algorithms, and Complex Algorithms

TL;DR

This paper extends Breiman's thesis by categorizing algorithmic modeling into parametric regressions, interpretable algorithms, and complex algorithms, analyzing their roles and differences across three cultures.

Contribution

It introduces a three-culture framework for algorithmic modeling, expanding Breiman's original two-culture thesis to include complex and interpretable algorithms.

Findings

Proposes a three-culture classification for modeling approaches.

Highlights differences between parametric, interpretable, and complex algorithms.

Discusses implications for statistical modeling and algorithm selection.

Abstract

We consider an extension of Leo Breiman's thesis from "Statistical Modeling: The Two Cultures" to include a bifurcation of algorithmic modeling, focusing on parametric regressions, interpretable algorithms, and complex (possibly explainable) algorithms.