Currently, there is no one common measure to judge machine learning methods and results. Through, this paper, I introduce such a measure to judge quality of machine learning methods and results. We just measure machine learning algorithms based on accuracy of prediction. If there is a common measure to define quality of machine learning methods and results then it could be used in different situations like interdisciplinary research. Moreover, the results are just limited to the data that we have in hand and any change in data changes the machine learning methods and results. Therefore, to become independent of data a measure is required to show the quality of methods in a way that is independent of data. This may be the first paper to present such a measure. When a measure independent of data is present, it will help us in determining the applicability of machine learning methods and results in a way that is independent of data. Correctness is such a measure. When I say correctness, I mean justified mathematically.

Machine Learning