Chocolate model with all coefficients random
Now that we have the effects coding stored with the chocolate data, we are ready to fit a model where all the coefficients are normally distributed. In order to do that, we need to create the rpar vector to input to mlogit(). That's a bit tricky, so I've written the code for you, but you should run it to see how it works. Then, you are going to write the call to mlogit().
Questo esercizio fa parte del corso
Choice Modeling for Marketing in R
Istruzioni dell'esercizio
- The first two inputs are the model formula
Selection ~ 0 + Brand + Type + Priceand the datachocolate. - The next input should be
rpar = my_rparwhich tellsmlogit()which coefficients we want to be normally distributed. - The last input should be
panel = TRUE.
Esercizio pratico interattivo
Prova a risolvere questo esercizio completando il codice di esempio.
# create my_rpar vector
choc_m2 <- mlogit(Selection ~ 0 + Brand + Type + Price, data=chocolate)
my_rpar <- rep("n", length(choc_m2$coef))
names(my_rpar) <- names(choc_m2$coef)
my_rpar
# fit model with random coefficients
choc_m7 <- mlogit(___)