A data-derived baseline using feature effects on binary outcomes provides a model-agnostic way to check if machine learning explanations align with the underlying data structure.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Does the Model Say What the Data Says? A Simple Heuristic for Model Data Alignment
A data-derived baseline using feature effects on binary outcomes provides a model-agnostic way to check if machine learning explanations align with the underlying data structure.