Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

Hit Rate in Latent Class Analysis


I know that there exist mixed recommendations regarding the calculation of hit rates for LC analysis. However, I would like to do so.
I've calculated the shares of preferences with the parameters of the LC model within the market simulator for each hold out incl. individual results. That gives me the probability for each respondent that he/she would choose concept 1 to 4. In order to compute the hits, is it correct to take the highest probability for one of the concepts/ respondent and then determine whether it matches the observed choice?
Is there any different way to conduct hit rates?

Best regards!
asked Jul 27, 2017 by briniminii (490 points)
retagged Jul 27, 2017 by Walter Williams

1 Answer

0 votes
As you have described is pretty much the way to do it - if the highest utility alternative matches the choice you have a hit, and if not then not.
answered Jul 27, 2017 by Keith Chrzan Platinum Sawtooth Software, Inc. (92,075 points)