ApolloChoiceModelling forum

Hi,

I'm new to Apollo and MDCEV.
I have trained an MDCEV model and tried to predict the choice.

I want to load the previous output, the "model" to do the prediction, like the following codes in the example

database = dataset1 # new dataset for prediction
apollo_inputs = apollo_validateInputs()
forecast_base = apollo_prediction(model,
apollo_probabilities,
apollo_inputs)

I noticed that although the estimated parameters are recorded in the model, the apollo_probabilities and the apollo_beta include initial settings for estimation, for example,
1. apollo_probabilities=function(apollo_beta, apollo_inputs, functionality="estimate"). Should I change the functionality to "prediction"?
2. a lot of settings in apollo_probabilities, such as #Define continuous consumption, are set for estimation, should I remove them?
3. the parameters in apollo_beta are the initial parameters before estimation, should I update it manually based on the results in the model?

Thank you so much for your time in advance!

Hi

you don't need to do any of those things. In prediction, you should use apollo_prediction, not call apollo_probabilities yourself. Have a look at the manual too where this is explained in detail

Stephane