Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{209426,
        author = {JEYA DAISY I and Dinesh Kumar V},
        title = {Golden Jackal Optimization-Driven Progressive Wasserstein Generative Adversarial Network for Wearable Cortisol-Based Stress Severity Classification},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {5},
        pages = {1758-1768},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=209426},
        abstract = {Chronic dysregulation of the hypothalamic-pituitary-adrenal axis is increasingly recognized as a driver of cognitive decline, cardiovascular disease and reduced quality of life, yet clinically meaningful, continuous, non-invasive tracking of stress remains an open problem. Sweat cortisol offers a biochemically direct proxy for circulating cortisol, but wearable sweat-sensing platforms still struggle with noisy, low-resolution readings, discontinuous flow and site-dependent variability, and the software pipelines built on top of them commonly rely on shallow classifiers whose hyper-parameters are chosen manually. This paper proposes SLM-CS-PWGAN, a stress-level monitoring pipeline that couples multi-site cortisol sensing with a Progressive Wasserstein Generative Adversarial Network (PWGAN) classifier whose hyper-parameters are tuned automatically by the Golden Jackal Optimization Algorithm (GJOA). Raw multi-site cortisol streams collected from eight participants are first denoised with an Unscented Trainable Kalman Filter, after which time-domain statistical descriptors are extracted through a Dual-Tree Complex Wavelet Transform. The resulting feature vectors are classified into Normal, Acute and Chronic stress categories by the PWGAN, whose generator and discriminator are grown progressively across resolutions to stabilize adversarial training, while GJOA searches the learning-rate, batch-size and architectural-depth space that conventional PWGAN training leaves to manual tuning. The paper details the complete architecture, the mathematical formulation of every stage, the evaluation protocol against three literature baselines, and the metrics on which the framework is to be benchmarked. The design is intended to reduce manual tuning effort and improve the stability of adversarial training relative to fixed-hyper-parameter PWGAN and the classical machine-learning baselines reviewed in this work.},
        keywords = {cortisol sensor, wearable stress monitoring, Golden Jackal Optimization, Progressive Wasserstein GAN, Kalman filtering, dual-tree complex wavelet transforms, hyper-parameter optimization},
        month = {October},
        }