EUVIMEDEUVIMED entdecken
Crossref

Explainable Machine Learning for Early Prediction of PCOS Using Clinical and Biochemical Features

Dr A R Jayasudha, E Ramya, B Raja, C Aswin Kumar, N Dhanush, D Dhanush

International Journal of Drug Delivery Technology · 30.05.2026 · 16 (38s)

DOI: 10.25258/ijddt.16.38s.52 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Drug Delivery Technology, online erschienen am 30.05.2026. Verfasst von Dr A R Jayasudha und 5 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The early detection of Polycystic Ovary Syndrome (PCOS) is critical for timely medical intervention and effective patient management. Traditional diagnostic methods rely on clinical examination, hormonal assays, and ultrasound imaging, which can be time-consuming and prone to human error. This paper presents a complete, reproducible pipeline for predicting PCOS using clinical, biochemical, and demographic features. We evaluate five supervised machine learning algorithms—Decision Tree, Random Forest, CatBoost, Logistic Regression, and Support Vector Machine (SVM)—on a structured patient dataset. The models achieved the following performance on the held-out test set: Decision Tree (Accuracy: 83.5%, F1-Score: 71.9%, AUC-ROC: 0.801), Random Forest (Accuracy: 87.2%, F1-Score: 76.7%, AUC-ROC: 0.945), CatBoost (Accuracy: 89.0%, F1-Score: 81.2%, AUC-ROC: 0.945), Logistic Regression (Accuracy: 85.3%, F1-Score: 73.3%, AUC-ROC: 0.838), and SVM (Accuracy: 71.6%, F1-Score: 6.1%, AUC-ROC: 0.244). A weighted scoring mechanism combining F1-Score (70%) and AUC-ROC (30%) identified CatBoost as the best-performing model. Explainability is provided through SHAP, highlighting clinically relevant attributes such as follicle count, LH/FSH ratio, BMI, and menstrual regularity. This reproducible framework demonstrates that machine learning can provide interpretable and reliable support for PCOS diagnosis.

Zitieren

APA

Jayasudha, D. A. R., Ramya, E., Raja, B., Kumar, C. A., Dhanush, N., & Dhanush, D. (2026). Explainable Machine Learning for Early Prediction of PCOS Using Clinical and Biochemical Features. International Journal of Drug Delivery Technology, 16(38s). https://doi.org/10.25258/ijddt.16.38s.52

Vancouver

Jayasudha DAR, Ramya E, Raja B, Kumar CA, Dhanush N, Dhanush D. Explainable Machine Learning for Early Prediction of PCOS Using Clinical and Biochemical Features. International Journal of Drug Delivery Technology. 2026;16(38s). doi:10.25258/ijddt.16.38s.52

BibTeX
@article{Jayasudha2026,
  author = {Jayasudha, Dr A R and Ramya, E and Raja, B and Kumar, C Aswin and Dhanush, N and Dhanush, D},
  title = {Explainable Machine Learning for Early Prediction of PCOS Using Clinical and Biochemical Features},
  journal = {International Journal of Drug Delivery Technology},
  year = {2026},
  volume = {16},
  number = {38s},
  doi = {10.25258/ijddt.16.38s.52}
}
Kostenlos starten →