ApolloChoiceModelling forum

I estimated a MNL model with weights (weights not equal to 1). Then, I'm hoping to apply the estimated model to the same dataset that I used in estimation, but got really wired output. What I did is simply using: predictions = apollo_predicition (model,apollo_probabilities, apollo_inputs) but got results like this

Predicted aggregated demand at model estimates
X0 X1 X2 X3 X4
Demand 0 0 0 3.18 0

I have 1222 observations in the estimation data and weighed sum is 338,225, so I'm not sure how to interpret the result of only 3.18 in aggregated demand. If I changed the weights to be 1s, the predicted aggregated demand output looks fine, but I'm hoping to predict with weights not equal to 1. Could anyone help with this? Thank you!

Hi,

Sorry for the long delay in the reply.

There has been some changes in how weights are used for prediction. I don't know what version you are using, but version 0.2.7 (currently on CRAN) should report both the weighted and unweighted prediction. Could you try again with the latest version of Apollo and let us know if that solved the problem?

Cheers
David