Rethinking Log Odds: Linear Probability Modelling and Expert Advice in Interpretable Machine Learning

TL;DR

This paper introduces Linearised Additive Models (LAMs) and SubscaleHedge, enhancing interpretability in machine learning models by replacing logistic links and combining expert advice without sacrificing performance.

Contribution

The paper presents LAMs and SubscaleHedge, novel methods that improve interpretability and combine models effectively in binary classification tasks.

Findings

LAMs provide direct probability attributions.

SubscaleHedge effectively combines feature subsets.

Algorithms maintain performance on financial data.

Abstract

We introduce a family of interpretable machine learning models, with two broad additions: Linearised Additive Models (LAMs) which replace the ubiquitous logistic link function in General Additive Models (GAMs); and SubscaleHedge, an expert advice algorithm for combining base models trained on subsets of features called subscales. LAMs can augment any additive binary classification model equipped with a sigmoid link function. Moreover, they afford direct global and local attributions of additive components to the model output in probability space. We argue that LAMs and SubscaleHedge improve the interpretability of their base algorithms. Using rigorous null-hypothesis significance testing on a broad suite of financial modelling data, we show that our algorithms do not suffer from large performance penalties in terms of ROC-AUC and calibration.