Extreme log-likelihood values get overflown when we attempt to sum their exponentials. However, using bigfloat package is an external library dependency that we can do without.
First course of action should be trying to calculate the e^ln(Z1) + e^ln(Z2) as a float128 type.
Extreme log-likelihood values get overflown when we attempt to sum their exponentials. However, using
bigfloat
package is an external library dependency that we can do without.First course of action should be trying to calculate the
e^ln(Z1) + e^ln(Z2)
as afloat128
type.