PR – Polychotomous Logistic Regression

Polychotomous Logistic Regression using BMDP Statistical Software Program PR

PR computes maximum likelihood estimates of logistic models for multinomial data. The categorical values of the dependent variables may be nominal or ordered. PR has the same capability of automatic generation of design variables as LR.

• Generalized version of LR that accommodates multinomial and ordinal dependent variables
• Stepwise MLE of the coefficients
• Step selection based on maximum likelihood ratio or asymptotic covariance estimate
• Log-likelihood and goodness-of-fit statistics
• Options to plot and save many diagnostic statistics
• View Program BMDP Program PR Manual Chapter