[Feature]: One step ahead (OSA) residuals for composition data
Rick-Methot-NOAA opened this issue · 0 comments
Rick-Methot-NOAA commented
Describe the solution you would like.
This paper describes a procedure for one step ahead residuals. It overcomes the shortcomings of Pearson residuals for composition data with multinomial logL
https://www.sciencedirect.com/science/article/pii/S0165783622002648?via%3Dihub
Describe alternatives you have considered
status quo
Statistical validity, if applicable
paper demonstrates superiority
Describe if this is needed for a management application
No response
Additional context
No response