Analogues of the Pearson correlation test

Zolotareva Tatyana

This paper considers discrete analogs of the Pearson correlation criterion, presents a numerical experiment that reproduces the conditions for modeling the spectral representation of data, and presents the results of a numerical experiment. The discrete-differential analogs of the Pearson correlation criterion are also considered. The dual representation of randomness using analogs of the Pearson correlation criterion is considered. Objective: to consider the dual representation of randomness in two different forms. Methods: discrete analogs of the Pearson correlation test and discrete-differential analogs of the Pearson correlation test. Results: the classical Pearson correlation coefficient and its differential-discrete analog are almost significantly less correlated (corr (r,"r13")-0.60871) than the Pearson difference correlations. At the same time, the probabilities of errors of the first and second kind for this class of transformations turn out to be applicable at a rate of 0.45 for practice. Conclusions: it is not the procedure of transition from one form of randomness description to another that is fundamentally important, but the fundamentally different properties of the data presented in two different forms.

Keywords:Pearson's criterion; discrete representations; continuous representations; neuron; error; continuum.