The Pythagorean fuzzy set (PFS) which is an expansion of intuitionistic fuzzy set, is more capable of expressing and handling the uncertainty under uncertain environment. In this paper, a novel Pythagorean information measure has been proposed along with its various properties for probability distributions and Pythagorean fuzzy sets. The new Pythagorean discriminant information measure has some desirable merits. We justify necessity of the newly proposed method using counterintuitive examples. To implement the application of Pythagorean measures in real life problem, we have taken real data from the repository of machine learning. After that we transform the real data set into PF-environment. Thereafter, by using the idea of degree of confidence the potential of proposed measure has been discussed. By contrasting the various methods it is found that the new divergence measure is more efficient as compared to other methods.