Criteria for Classifying Forecasting Methods

TL;DR

This paper critiques the common classification of forecasting methods into machine learning or statistical categories, arguing that this distinction is superficial and proposing alternative criteria for meaningful comparison and collaboration.

Contribution

It challenges the traditional dichotomy in forecasting classification and offers new characteristics to better understand and evaluate forecasting methods.

Findings

The ML/statistical distinction is more tribal than fundamental.

Alternative characteristics can improve method classification.

Cross-pollination between ML and statistical forecasting can enhance effectiveness.

Abstract

Classifying forecasting methods as being either of a "machine learning" or "statistical" nature has become commonplace in parts of the forecasting literature and community, as exemplified by the M4 competition and the conclusion drawn by the organizers. We argue that this distinction does not stem from fundamental differences in the methods assigned to either class. Instead, this distinction is probably of a tribal nature, which limits the insights into the appropriateness and effectiveness of different forecasting methods. We provide alternative characteristics of forecasting methods which, in our view, allow to draw meaningful conclusions. Further, we discuss areas of forecasting which could benefit most from cross-pollination between the ML and the statistics communities.