Predictive performance

Building Response Models in R

Kathrin Gruber

Assistant Professor of Econometrics Erasmus University Rotterdam

Classification

predicted <- ifelse(fitted.values(extended.mod) >= 0.5, 1, 0)

table(predicted)/2798

predicted
      0       1 
0.96605 0.03395

observed <- choice.data$HOPPINESS
table(observed, predicted)/2798

        predicted
observed       0       1
       0 0.88849 0.01144
       1 0.07756 0.02252

library(pROC)
Roc <- roc(predictor = fitted.values(extended.mod), response = observed)

plot(Roc)

Building Response Models in R