Model averaging approach

This forum is for posts that specifically focus on Ngene.

Moderators: Andrew Collins, Michiel Bliemer, johnr

Model averaging approach

Postby psalazar » Thu Oct 20, 2022 6:01 pm

Dear professor Bliemer,

I am using the model averaging approach, following the example from the Ngene manual (page 132). If I have two models (M1 and M2), and I multiply the efficiency measure for model ‘M1’ by 1.5 in the Ngene syntax, would that mean that the weight I'm giving to M1 is 0.6 and the weight of M2 would be 0.4? I read your paper with Rose and Scarpa (2009) on the model averaging approach (great paper and very easy to follow!), and in there, you present the case study with the 5 models, using three different weights (table 2), but I'm not sure about the Ngene syntax equivalent. For the example in the Ngene manual (page 133), would this be correct?

M1 --> In the Ngene syntax (eff): 1 ; weight: 0.15
M2 --> In the Ngene syntax (eff): 2 ; weight: 0.31
M3 --> In the Ngene syntax (eff): 1.5 ; weight: 0.23
M4 --> In the Ngene syntax (eff): 1 ; weight: 0.15
M5 --> In the Ngene syntax (eff): 1 ; weight: 0.15

Many thanks!

Best wishes,

Pamela.
psalazar
 
Posts: 15
Joined: Fri Jun 24, 2022 1:42 am

Re: Model averaging approach

Postby Michiel Bliemer » Tue Nov 01, 2022 9:50 am

Sorry for the delayed response, I was on holidays for a week :)

The weight is not exactly the weight of a MODEL, but rather a weight of the EFFICIENCY CRITERION for that model.

For example:

;eff = 2* m1(mnl,d) + 1* m2(mnl,a)

Then the combined efficiency criterion that is minimised is 2 times the D-error of the design for model m1 and 1 time the A-error of the design for model m2.

Michiel
Michiel Bliemer
 
Posts: 1705
Joined: Tue Mar 31, 2009 4:13 pm


Return to Choice experiments - Ngene

Who is online

Users browsing this forum: No registered users and 11 guests