Balancing model complexity and generalisation: Effect of ANN depth on surface roughness prediction in turning of SUS 304
Journal of Science and Development Economics · 01.07.2026 · (42)
DOI: 10.64632/jsde.42.2026.849 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Science and Development Economics, online erschienen am 01.07.2026. Verfasst von Huynh Thanh Thuong und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
This study examines how the depth of Artificial Neural Network (ANN) architectures affects the prediction of surface roughness (Ra) in finish turning of SUS 304 stainless steel. Experimental data were generated using a Taguchi L27 design with cutting speed, feed rate, tool nose radius, and machining diameter as inputs. ANN models with one to four hidden layers were developed and evaluated through a unified framework including training, internal testing, and external validation, supported by 12 additional experiments beyond the design space. Results show that ANN depth significantly impacts predictive performance, but increasing complexity does not guarantee better accuracy under limited data conditions. The two-hidden-layer ANN achieved the best balance between accuracy and generalisation, outperforming both simpler and deeper models. These findings offer clear, practical guidance for selecting appropriate ANN architectures in machining applications, particularly under limited experimental data.
Zitieren
APA
Thuong, H. T., & Huynh, T. P. (2026). Balancing model complexity and generalisation: Effect of ANN depth on surface roughness prediction in turning of SUS 304. Journal of Science and Development Economics. https://doi.org/10.64632/jsde.42.2026.849
Vancouver
Thuong HT, Huynh TP. Balancing model complexity and generalisation: Effect of ANN depth on surface roughness prediction in turning of SUS 304. Journal of Science and Development Economics. 2026. doi:10.64632/jsde.42.2026.849
BibTeX
@article{Thuong2026,
author = {Thuong, Huynh Thanh and Huynh, Thanh Phong},
title = {Balancing model complexity and generalisation: Effect of ANN depth on surface roughness prediction in turning of SUS 304},
journal = {Journal of Science and Development Economics},
year = {2026},
number = {42},
doi = {10.64632/jsde.42.2026.849}
}
EUVIMED