Machine Learning Innovations in Polycystic Ovarian Syndrome Diagnosis: A Comprehensive Review

Polycystic Ovarian Syndrome (PCOS) is a non-fatal yet serious ailment as it is a precursor to other serious and life threatening diseases like cancer, diabetes, etc. It is also a leading cause of infertility among women as it affects around 13% of the women after the attainment of reproductive age. The exact causes of PCOS remain unknown; however, common clinical symptoms are usually easy to diagnose. Many states of the art machine learning methodologies have been developed in order to diagnose PCOS autonomously without the intervention of doctors as it helps reduce the patient load and improves the efficiency of the medical professional. The following paper reviews such latest developments and helps get a better understanding of the various steps that can be taken in order to improve the sustaining methodologies. The current scenario and the reliability of the various machine learning models stand up to decision support systems. The paper also talks about various image augmentation and image processing techniques along with multiple ensemble methods in order to decrease the computational time and complexity without trading off for the performance metrics. The diagnosis of PCOS at the right time is extremely crucial as for the various ailments associated with it. It can increase the risk for mental agony and higher risk for suicides. The basic step in order to detect PCOS is selection of follicles which can be classified in order to get the results.