Convert to Positive Definite/Semi-Definite - Asset Allocation issue

Hi:

I need some help in identifying a procedure, better yet the algorithm
:), to convert a correlation matrix that is not positive
semi-definite into its "closest" non-negative form.

The problem has arisen because we are using the same correlations for
two asset classes. The inelegant method is to vary/perturb the
correlations of one of the two offending asset classes by a small
fraction. This works. However, I am looking for a more elegant
solution that I can implement, which does not require the matrix to
be eye-balled.

Thanks.

Faiz
.

Relevant Pages

  • Re: Convert to Positive Definite/Semi-Definite - Asset Allocation issue
    ... Faiz Hakim wrote: ... to convert a correlation matrix that is not positive ... > The problem has arisen because we are using the same correlations ... > correlations of one of the two offending asset classes by a small ...
    (comp.soft-sys.matlab)
  • Re: Convert to Positive Definite/Semi-Definite - Asset Allocation issue
    ... Phil Goddard wrote: ... to convert a correlation matrix that is not positive ... >> two asset classes. ... >> correlations of one of the two offending asset classes by a ...
    (comp.soft-sys.matlab)
  • Re: PCA C code contradictions
    ... how I advised simple correlations to be used. ... the formula for the REAL correlation matrix! ... Most typos are inconsequential, as the typo in the title above. ... that betray the claim of the misspeller as an expert in statistics. ...
    (sci.stat.math)
  • Re: efficient computation of many univariate least squares fits
    ... The suggestion in a different subthread to use the master correlation matrix may be a good alternative, particularly if you're using something that has a built-in routine for calculating correlations and it has a built-in t-test for the significance of the correlations. ...
    (sci.stat.math)
  • newbie multivariate question
    ... i can report the correlation matrix. ... i have no reason to believe that ... correlations between a, c and d ...
    (sci.stat.consult)
Loading