combining a multinomial and conditional logit model

For questions regarding programming in the EViews programming language.

Moderators: EViews Gareth, EViews Jason, EViews Moderator, EViews Matt

jejafo
Posts: 2
Joined: Tue Apr 19, 2016 12:22 pm

combining a multinomial and conditional logit model

Postby jejafo » Thu May 25, 2017 4:26 am

Hello,

For my research I want to estimate parameters for a unordered logit model, where some of parameters should vary across the different options of the dependent variable (j=1,..7) and some of the parameters should be the same for each option. I added a simple example of my model to illustrate my question with j=2 and two variables (+ one constant).

xb1 = b_0,1 + b_1*price1 + b_2,1*income
xb2 = b_0,2 + b_1*price2 + b_2,2*income
xb3 = [0,0,0]
LOGL1=BRAND1*EXP(XB1)+BRAND2*EXP(XB2)+BRAND3*EXP(XB3)-LOG(DENOM)
DENOM=1+EXP(XB1) + EXP(XB2)

In this case the b_1 coefficient should be estimated by means of Conditional logit, while the b_2 coefficents should be estimated by Multinomial logit. I saw guides how to do both of them, but nowhere on internet I could find any instructions how to combine the two. I was hoping someone could help me out.

Thanks in advance!

Return to “Programming”

Who is online

Users browsing this forum: No registered users and 30 guests