Endocrine Disorder Classification Using Machine Learning Models Polycystic Ovary Syndrome Diagnosis

Polycystic ovary syndrome (PCOS) is one of the most common of endocrine disorders in women of reproductive age characterized by oligomenorrhea, hyperandrogenism, and polycystic ovarian morphology. Standard procedures to diagnose PID mainly depend on ultrasound imaging and hormonal assays which are affected by issues of manual interpretation, variability and delayed results. An automated PCOS detection and classification decision support system based on important hormonal and metabolic indicators (LH, FSH, insulin, testosterone, glucose and BMI). The system consist of normalisation, imputation as data preprocessing, Then features extraction are performed during diagnosis. In this work, we use advanced machine learning algorithms XGBoost, Random Forest (RF) and Support Vector Machine (SVM) to classify subjects as having PCOS (all types) or not PCOS as well as subtypes of PCOS including hyperandrogenic PCOS, ovulatory PCOS and metabolic PCOS. The results of the prediction and risk probabilities are used to guide health recommendations with a user-friendly diagnostic interface and the potential for implementation in a clinical setting. The framework is a combination of AI and predictive analytics, reducing the variabilities in diagnosis, that is, it avoids human biases and can be scaled widely through screening via hospitals and telehealth networks. This AI based tool for PCOS detection ensures early identification, timely intervention and better reproductive-health in women.