Publication Date



Technical Report: UTEP-CS-20-107


Data for processing mostly comes from measurements, and measurements are never absolutely accurate: there is always the "measurement error" -- the difference between the measurement result and the actual (unknown) value of the measured quantity. In many applications, it is important to find out how these measurement errors affect the accuracy of the result of data processing. Traditional data processing techniques implicitly assume that we know the probability distributions. In many practical situations, however, we only have partial information about these distributions. In some cases, all we know is the upper bound on the absolute value of the measurement error. In other cases, data comes not from measurements but from expert estimates. In this paper, we explain how to estimate the accuracy of the results of data processing in all these situations. We tried to explain not only what methods can be used, but also why these methods have been proposed and have been successfully used. We hope that this overview will be helpful both to users solving practical problems and to researchers interested in extending and improving the existing techniques.