Output layer configurations for binary classification: single-neuron versus dual-neuron approaches
PeerJ Computer Science · 13.07.2026 · 12 · e4017
DOI: 10.7717/peerj-cs.4017 ↗Kurz gefasst
Fachartikel in der Zeitschrift PeerJ Computer Science, online erschienen am 13.07.2026. Verfasst von Khadim Hussain und 4 weiteren Autor:innen. Die Arbeit nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Science and Technology Council (NSTC), Taiwan, Chang Gung University and Chang Gung Memorial Hospital.
Abstract (Originalsprache)
Binary classification is a fundamental machine learning problem with applications spanning image recognition, medical diagnosis, and natural language processing. Neural networks typically use either a single output neuron with sigmoid activation or two output neurons with softmax activation to model class probabilities. However, practical guidance on choosing between these output configurations remains unclear in many real-world scenarios. This study provides a comprehensive empirical comparison of these two output layer configurations, investigating their influence on model performance, training dynamics, and generalization capabilities. Despite the theoretical equivalence of single-neuron and dual-neuron output layers for binary classification, there is a notable lack of systematic empirical studies comparing their impact on model performance, convergence, and generalization. This absence of clear, data-driven guidance poses challenges for practitioners, especially as neural network architectures grow more complex. Through extensive experimentation on diverse neural network architectures—including a custom Small Convolutional Neural Network (CNN), Vision Transformer (ViT), and ResNet-50—applied to benchmark datasets derived from CIFAR-10, we identify conditions under which each approach shows superior characteristics. Both output layer configurations were implemented across all architectures and trained under identical protocols. In these standard implementations, single-neuron models use binary cross-entropy (BCE) and dual-neuron models use categorical cross-entropy (CCE); therefore, observed differences reflect this paired design. Our findings are based on five-seed repeats per setting (mean ± std) and per-task paired statistical tests. The single-neuron output with sigmoid activation and BCE loss shows a statistically significant advantage for the Small CNN on an easier task, while modern architectures (ViT, ResNet-50) show comparable performance with small, non-significant trends, with absolute gaps typically
Förderung
- National Science and Technology Council (NSTC), Taiwan
- Chang Gung University and Chang Gung Memorial Hospital
Zitieren
APA
Hussain, K., Khan, T. F., Hassan, A., Khan, L., & Chang, H. (2026). Output layer configurations for binary classification: single-neuron versus dual-neuron approaches. PeerJ Computer Science, 12, e4017. https://doi.org/10.7717/peerj-cs.4017
Vancouver
Hussain K, Khan TF, Hassan A, Khan L, Chang H. Output layer configurations for binary classification: single-neuron versus dual-neuron approaches. PeerJ Computer Science. 2026;12:e4017. doi:10.7717/peerj-cs.4017
BibTeX
@article{Hussain2026,
author = {Hussain, Khadim and Khan, Talha Farooq and Hassan, Ali and Khan, Lal and Chang, Hsien-Tsung},
title = {Output layer configurations for binary classification: single-neuron versus dual-neuron approaches},
journal = {PeerJ Computer Science},
year = {2026},
volume = {12},
pages = {e4017},
doi = {10.7717/peerj-cs.4017}
}
EUVIMED