Certain investigations on prognosis of polycystic ovarian syndrome using machine learning approaches for early treatment

Abstract

Polycystic ovary syndrome (PCOS) is a metabolic and endocrine disorder that affects women of childbearing age. This syndrome is classified by polycystic ovaries, irregular cycles, and hyperandrogenism. The PCOS etiology is not so clear; however, it is considered to as multi-factorial. It provides a stronger relationship among hyper-androgenism and hyper-insulinemia, but the factors behind this association cannot be predicted. Abnormal metabolic and obesity profiles are common with PCOS women, and 60% - 70% are insulin resistant, which increases the risk of Type II diabetes, without considering age and Body Mass Index (BMI). The women with PCOS shows reduced fertility rate, and during the pregnancy either by assisted reproduction techniques or by natural process, they are supposed to have higher risk in complications like diabetes or pre-eclampsia that aggravates the fetus health and the health of the women. Sometimes, it leads to intra-uterine condition that affects the fetus. In women of PCOS, there are higher chance of fetus get affected hypothetically by direct exposure towards maternal androgens or through placenta dysregulation. For both the pregnant and non-pregnant women of PCOS, the modification of lifestyle with physical exercise and diet are the primary level of treatment. However, to avoid the scary condition after the prediction of PCOS has to be prevented with beneficial outcomes. To address the above issues, this investigation concentrates on modelling a predictor approach to identify the factors that influences PCOS in women at earlier stage using newline