Improving the explainability of autoencoder factors for commodities through forecast-based Shapley values

Abstract Autoencoders are dimension reduction models in the field of machine learning which can be thought of as a neural network counterpart of principal components analysis (PCA). Due to their flexibility and good performance, autoencoders have been recently used for estimating nonlinear factor mo...

Full description

Bibliographic Details

Published in:	Scientific Reports
Main Authors:	Roy Cerqueti, Antonio Iovanella, Raffaele Mattera, Saverio Storani
Format:	Article
Language:	English
Published:	Nature Portfolio 2024-08-01
Subjects:	Explainability Neural networks Nonlinear factor models Shapley value Commodities
Online Access:	https://doi.org/10.1038/s41598-024-70342-5