We also demonstrated a technique for generating correlation matrices using uniform random variable distribution within the boundaries of each correlation coefficient. The theoretical bounds are sequentially computed.
Numpacharoen, Kawee and Atsawarungruangkit, Amporn, Generating correlation matrices based on the boundaries of their coefficients (September 20, 2012). Available at SSRN: http://ssrn.com/abstract=2127689
Kawee Numpacharoen (2021). Generating correlation matrices based on the boundaries of their coefficients. (https://www.mathworks.com/matlabcentral/fileexchange/37804-generating-correlation-matrices-based-on-the-boundaries-of-their-coefficients), MATLAB Central File Exchange. Retrieved .
the following command seems superior to this file
An important restriction is that the nondiagonal elements of the inverse correlation matrix are zero, except for one column and row.
What is the meaning of the input K? Does it somehow affect the minimal eigenvalue p? (I cannot download the paper in SSRN.)
Find the treasures in MATLAB Central and discover how the community can help you!Start Hunting!