Research Information System
3rd International Conference on Mathematics and its Applications in Science and Engineering, ICMASE 2022, Bucharest, Romania, 4 - 07 July 2022, vol.414, pp.273-284
The covariance and correlation are important indicators to measure the linear dependence between two variables. The multivariables as p-variables in the research are commonly observed in the engineering, social science, medical examination, etc. when p-variables have negative and positive dependence. In this study, the definition of correlation coefficient is reflected to define a multivariate generalization for the correlation. The linear dependence among p-variables are generated by using the artificial data sets which have the linear dependence and normal distribution. The different sample sizes and different number of p-variables are used in the simulation study. Even if the number of p-variables is greater than the number of sample size, the dependence coefficient shows better performance. Thus, the numerical results have shown that the proposed multicorrelation coefficient shows better performance to detect the value of degree of dependence among p-variables.