I recently ran the same data set through the SSAS logistic regression (Neural network - Hidden layer disabled) as well as a vanilla logistic regression procedure. I am aware that different processes are followed in order to obtain the resulting co-efficients,
however, has anyone got further information as to how close the results of the SSAS logistic regression equation are in comparison to logistic regression equations contained in other statistical packages such as SAS or SPSS?
I do understand that the respective procedures are dependant on the quality of the data that they are applied to. I assume that the SSAS logistic regression may perform more favourably on one dataset in comparison to a vanilla logistic regression and vice
I am aware of the following post however it does not go into much detail concerning the results of the comparison.
Further info would certainly be appreciated.
View Complete Post