penalty0 {MQMF}

penalty0 enables the adding of a large penalty as one approaches 0.0

Description

penalty0 allows for the option of adding a large penalty as a parameter approaches 0.0 . See negLL1 for example code that contains such a parameter. For example, when fitting an spm sometimes the optimal mathematical model fit can occur by depressing the r value to 0 or even go negative. Input values < 0.006 begin to generate large values as one goes smaller. The examples below illustrate this.

Usage

penalty0(x)

Arguments

x

the parameter value that potentially incurs a penalty

Value

a single value as a penalty to be added to a Log-Likelihood or SSQ

Examples

  penalty0(0.5)
  penalty0(0.1)
  penalty0(0.01)
  penalty0(0.005)