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

Abstract

Understanding the data and reaching accurate conclusions are of paramount importance in the present era of big data. Machine learning and probability theory methods have been widely used for this purpose in various fields. One critically important yet less explored aspect is capturing and analyzing uncertainties in the data and model. Proper quantification of uncertainty helps to provide valuable information to obtain accurate diagnosis. This paper reviewed related studies conducted in the last 30 years (from 1991 to 2020) in handling uncertainties in medical data using probability theory and machine learning techniques. Medical data is more prone to uncertainty due to the presence of noise in the data. So, it is very important to have clean medical data without any noise to get accurate diagnosis. The sources of noise in the medical data need to be known to address this issue. Based on the medical data obtained by the physician, diagnosis of disease, and treatment plan are prescribed. Hence, the uncertainty is growing in healthcare and there is limited knowledge to address these problems. Our findings indicate that there are few challenges to be addressed in handling the uncertainty in medical raw data and new models. In this work, we have summarized various methods employed to overcome this problem. Nowadays, various novel deep learning techniques have been proposed to deal with such uncertainties and improve the performance in decision making.

Keywords: Bayesian inference; Classification; Fuzzy systems; Machine learning; Monte Carlo simulation; Uncertainty.

Fig. 1

Having uncertainty makes the decision-making process difficult

Fig. 2

Paper selection mechanism

Fig. 3

Number of papers published in handling uncentainity in medical data between 1991 and 2020

Fig. 4

Most common algorithms used to handle uncertainties

Fig. 5

Classical backpropagation sets fixed values as weights (left) while in Bayesian neural network a distribution is assigned to each weight (right)

Fig. 6

Type-3 ANFIS structure

Fig. 7

Three disjoint regions of a rough set

Fig. 8

Percentage of uncertainty handling algorithms used in medical field researches

Fig. 9

The trade-off between data and model uncertainty