Handling of uncertainty in medical data using machine learning and probability theory techniques: A review of 30 years (1991-2020)

Understanding data and reaching valid conclusions are of paramount importance\nin the present era of big data. Machine learning and probability theory methods\nhave widespread application for this purpose in different fields. One\ncritically important yet less explored aspect is how data and model\nuncertainties are captured and analyzed. Proper quantification of uncertainty\nprovides valuable information for optimal decision making. This paper reviewed\nrelated studies conducted in the last 30 years (from 1991 to 2020) in handling\nuncertainties in medical data using probability theory and machine learning\ntechniques. Medical data is more prone to uncertainty due to the presence of\nnoise in the data. So, it is very important to have clean medical data without\nany noise to get accurate diagnosis. The sources of noise in the medical data\nneed to be known to address this issue. Based on the medical data obtained by\nthe physician, diagnosis of disease, and treatment plan are prescribed. Hence,\nthe uncertainty is growing in healthcare and there is limited knowledge to\naddress these problems. We have little knowledge about the optimal treatment\nmethods as there are many sources of uncertainty in medical science. Our\nfindings indicate that there are few challenges to be addressed in handling the\nuncertainty in medical raw data and new models. In this work, we have\nsummarized various methods employed to overcome this problem. Nowadays,\napplication of novel deep learning techniques to deal such uncertainties have\nsignificantly increased.\n