`PsiMetalog(min, max, coefficients, prop_fcns)`

A general meta-log distribution is defined in models through PsiMetalog(). This function has as arguments a vector of coefficients (coefficients) and optionally lower (min) and upper (max) bounds. Since the coefficients do not have a visible relation to reality, we must compute them from a set of historical data pairs {y, x}, where y is a cumulative probability and x is the correspondent percentile. The coefficients argument is either a range or an array of 2 to 10 numerical elements.

There are 3 optional cases for this distribution:

- Both
*min*and*max*are missing which results in an unbounded distribution Example:(, , {33, 3.7, -2.4, 5.7, 10.1})*PsiMetalog* - Only
*min*exists which results in a semi-bounded distribution Example:(0, , {33, 3.7, -2.4, 5.7, 10.1})*PsiMetalog* - Both
*min*and*max*exist which results in a bounded distribution Example:(0, 55, {33, 3.7, -2.4, 5.7, 10.1})*PsiMetalog*

Note: An exclusive bounding using only the max argument is not supported. The min and max arguments are true lower and upper bounds not to be confused with Lower Truncate and Upper Truncate. Optional standard property functions may be used with this distribution, i.e. PsiTruncate, PsiShift, PsiCorrelate, etc. The distribution may also be compounded.

Analytic moments are available only for unbound distributions. Kurtosis is not computed for more than 5 coefficients.