Estimation of Correlation Coefficients from Scatter Diagrams

Abstract

Theoretical relations between statistical scatter diagrams of two random variables and their joint probability density functions are used to derive a simple method of estimating correlation coefficients. The cases considered are those of variables with Gaussian and Rayleigh distributions. Theoretical curves and experimental scatter diagrams are given for each case. Some factors which may influence the accuracy of the proposed method are discussed, and it is concluded that the errors in estimating correlation coefficient will be of the order ±0.1.