ApolloChoiceModelling forum

Hi

We are estimating a model in wtp-space. The estimated parameters are sensitive to starting values, and we need to run robustness checks varying the starting values. Is there a command in apollo that automatically repeats the estimation of a model updating the starting values of the parameters?

We are aware of the apollo_searchStart command, but not sure that this solve our problem. What we need is, after having searched for and found the best starting values, a command that can verify that these starting values are producing a global optimum by testing for small deviations. And if we are not able to verify a global optimum, we need at least to check that we have a stable local optimum.

thank you in advance for your reply.

Best regards,
Margrethe

Margrethe

you can never know whether you found the global optimum as you cannot test all possible solutions.

What you could do re the below is write a loop with lots of starting values and carry out the full estimation from each of these

Stephane