id	author	title	date	pages	extension	mime	words	sentence	flesch	summary	cache	txt
bracis-19050	Silva Filho, Rogério Luiz Cardoso; Adeodato, Paulo Jorge Leitão; Brito, Kellyton dos Santos	Interpreting Classification Models Using Feature Importance Based on Marginal Local Effects	2021		.htm	text/html	5831	322	49	Feature importance (MUA) by groups during the period (RF and AB means) Full size image 5 Discussion and Conclusion This paper has proposed new model-agnostic metrics of feature importance in an attempt to circumvent the drawbacks and constraints of the existing methods, such as the β- coefficients of additive models and feature importance from tree-based algorithms, widely used for this purpose. In this paper, the main goal of this paper is to support applied research providing single metrics that, in a more realistic scenario, are able to report the overall contribution of model features.	cache/bracis-19050.htm	txt/bracis-19050.txt