PCOS PREDICTION & DETECTION IN WOMEN USING MACHINE LEARNING ALGORITHMS

Abstract

Pre-term abortions, infertility, and anovulation are all common problems among today's women.Infertility has been linked significantly to the disorder polycystic ovarian syndrome (PCOS), which affects women of reproductive age. Women who are in their reproductive years are affected by this hormonal disease. The hormonal imbalance in women  causes menstrual periods that are late or non-existent. Excessive weight gain, facial hair development, acne, hair loss, skin discoloration, pregnancy concerns, and irregular periods are all things to be aware of. are typical signs of PCOS, which in some women can cause infertility. The syndrome cannot be cured after it has been identified, however proper treatment can help ease its symptoms. Due to the wide range of symptoms and the existence of a variety of associated gynecological problems, PCOS is exceedingly difficult to diagnose.The energy and cost invested on a slew of clinical testing and ovary scanning has become a hardship for PCOS patients. To solve this problem, we suggest a method based on an ideal and basic set of characteristics that can help in the early identification and prediction of PCOS treatment. Machine learning classifiers such as Random Forest, SVM, Logistic Regression, Gaussian Naive Bayes, and K Neighbors were used to predict the underlying state of PCOS in women. The CHI SQUARE method was used to determine the top 30 features from the dataset, which were then included in the feature vector. The Random Forest Classifier offers the best and most dependable accuracy, according to our comparison of the findings of each classifier. Prasoon Kottarathil owns the dataset used for training and testing, which is freely accessible on KAGGLE