Hi!
First of all, thanks for your package. It is extremely easy to implement and super useful!
I have a question regarding H-statistic computation. I read that the limitation of H-statistic is that the normalized values [0-1], which are the standard values, only account for the relative strength of the interaction but hide if the interaction was relevant for the model on the big scale.
I wonder if there is a way to compute unnormalized H-statistic and plot them using IML. The tutorial essentailly says that normalized values are used but does not suggest any alternative.
All best and keep up the good work,
Jose
Hi!
First of all, thanks for your package. It is extremely easy to implement and super useful!
I have a question regarding H-statistic computation. I read that the limitation of H-statistic is that the normalized values [0-1], which are the standard values, only account for the relative strength of the interaction but hide if the interaction was relevant for the model on the big scale.
I wonder if there is a way to compute unnormalized H-statistic and plot them using IML. The tutorial essentailly says that normalized values are used but does not suggest any alternative.
All best and keep up the good work,
Jose