A prediction model for poly-cystic ovary syndrome problem using computational intelligence

Abstract

Women of childbearing age may face infertility owing to polycystic ovarian syndrome (PCOS).This illness causes ovarian dysfunction, which increases the chance of miscarriage and stillbirth, thus early management is necessary for a healthy lifestyle and to avoid future infections. Weight gain, irregular menstruation cycles, thinning hair, acne, dark and thick spots on the back of the neck, and anxiety disorders are the main symptoms of PCOS. A single out of five women has PCOS. Often women ignore the common symptoms of PCOS and wait until pregnancy problems emerge to get care. Considering PCOS is associated with an increased risk of developing a number of illnesses, such as glucose intolerance, , elevated cholesterol levels, and cardiovascular diseases, it should be identified as soon as feasible. The current tools and therapies are insufficient to identify and forecast PCOS at an earlier stage. To address this issue, we developed a model that will aid in the early detection of PCOS utilizing machine learning techniques and an absolute and minimal set of parameters. The Extra Tree Classifier, a forward selection approach followed by Wrapper, the Chi-square test, and Pearson Correlation was employed as selection criteria to evaluate essential characteristics. KAGGLE has a database that is used for training and testing.