KrishnaMohan
09-12-2005, 08:29 AM
Dear list
i have some 15 X 15 matrices - which are Hessian , and they are suppose to be Symmetric, generated from some quantum mechanical packages.
the problem is - All these matrices are NON SYMMETRIC.
Since i need to Quantify the OVERALL - ERROR of these matrices with respect to THE IDEAL Hessian that i already have, HOW CAN I DO THAT ?
OF COURSE I can compare EACH AND EVERY element of the matrices and that of ideal Hessian and take a difference, or just compare the DIAGONAL ELEMENT DIFFERENCE / DETERMINENTS of these matrices and ideal.
BUT I REALLY NEED A SINGLE-VALUED QUANTITY TO MEASURE IT.
I CHECKED THE N.R. IN C, but i did not find any info abt this-explicitly.
DOES ANYBODY HELP ME TO SOLVE THIS PROBLEM.
[I THINK SOMEBODY HAVE A SUFFICIENT STATISTICAL DATA ANALYSIS CAN GIVE A SHORT ANSWER EASILY ...!]
THANKS & CHEERS
KRISHNA
i have some 15 X 15 matrices - which are Hessian , and they are suppose to be Symmetric, generated from some quantum mechanical packages.
the problem is - All these matrices are NON SYMMETRIC.
Since i need to Quantify the OVERALL - ERROR of these matrices with respect to THE IDEAL Hessian that i already have, HOW CAN I DO THAT ?
OF COURSE I can compare EACH AND EVERY element of the matrices and that of ideal Hessian and take a difference, or just compare the DIAGONAL ELEMENT DIFFERENCE / DETERMINENTS of these matrices and ideal.
BUT I REALLY NEED A SINGLE-VALUED QUANTITY TO MEASURE IT.
I CHECKED THE N.R. IN C, but i did not find any info abt this-explicitly.
DOES ANYBODY HELP ME TO SOLVE THIS PROBLEM.
[I THINK SOMEBODY HAVE A SUFFICIENT STATISTICAL DATA ANALYSIS CAN GIVE A SHORT ANSWER EASILY ...!]
THANKS & CHEERS
KRISHNA