Superior Scoring Rules for Probabilistic Evaluation of Single-Label Multi-Class Classification Tasks

TL;DR

This paper proposes novel scoring rules, PBS and PLL, that improve probabilistic model evaluation by better rewarding correct classifications, leading to enhanced model selection and higher F1 scores.

Contribution

Introduction of Penalized Brier Score and Penalized Logarithmic Loss, novel proper scoring rules that incorporate penalties for misclassifications to improve model evaluation.

Findings

PBS correlates better with F1 score than Brier Score.

Models selected by PBS and PLL achieve higher F1 scores.

PBS improves early stopping and checkpointing decisions.

Abstract

This study introduces novel superior scoring rules called Penalized Brier Score (PBS) and Penalized Logarithmic Loss (PLL) to improve model evaluation for probabilistic classification. Traditional scoring rules like Brier Score and Logarithmic Loss sometimes assign better scores to misclassifications in comparison with correct classifications. This discrepancy from the actual preference for rewarding correct classifications can lead to suboptimal model selection. By integrating penalties for misclassifications, PBS and PLL modify traditional proper scoring rules to consistently assign better scores to correct predictions. Formal proofs demonstrate that PBS and PLL satisfy strictly proper scoring rule properties while also preferentially rewarding accurate classifications. Experiments showcase the benefits of using PBS and PLL for model selection, model checkpointing, and early stopping. PBS exhibits a higher negative correlation with the F1 score compared to the Brier Score during training. Thus, PBS more effectively identifies optimal checkpoints and early stopping points, leading to improved F1 scores. Comparative analysis verifies models selected by PBS and PLL achieve superior F1 scores. Therefore, PBS and PLL address the gap between uncertainty quantification and accuracy maximization by encapsulating both proper scoring principles and explicit preference for true classifications. The proposed metrics can enhance model evaluation and selection for reliable probabilistic classification.