Page 1 of 1

Hybrid model with OL

Posted: 03 Apr 2022, 20:40
by Yashin
Dear Concern,

I am trying to estimate the the Hybrid model with OL, but I cam across the section of the code which I don't understand,

"choiceAnalysis_settings <- list(
alternatives = c(Artemis=11, Novum=12, BestValue=21, Supermarket=22, PainAway=23),
avail = with(database,
list(Artemis = brand_1=="Artemis" | brand_2=="Artemis" ,
Novum = brand_1=="Novum" | brand_2=="Novum" ,
BestValue = brand_3=="BestValue" | brand_4=="BestValue" ,
Supermarket = brand_3=="Supermarket" | brand_4=="Supermarket",
PainAway = brand_3=="PainAway" | brand_4=="PainAway" )
),
choiceVar = with(database,
11*((best==1 & brand_1=="Artemis" ) | (best==2 & brand_2=="Artemis" )) +
12*((best==1 & brand_1=="Novum" ) | (best==2 & brand_2=="Novum" )) +
21*((best==3 & brand_3=="BestValue" ) | (best==4 & brand_4=="BestValue" )) +
22*((best==3 & brand_3=="Supermarket") | (best==4 & brand_4=="Supermarket")) +
23*((best==3 & brand_3=="PainAway" ) | (best==4 & brand_4=="PainAway" ))),
explanators = database[,c("regular_user","university_educated","over_50")]
)

apollo_choiceAnalysis(choiceAnalysis_settings, apollo_control, database)"

What does 11, 12 ,13 mean? Since, these values have been multiplied by the brand values I did not understand properly.

Any suggestions are highly appreciated.

Re: Hybrid model with OL

Posted: 04 Apr 2022, 17:17
by stephanehess
Hi

this is not really related to the model specification/estimation, but to pre-analysis of the data, as described in the manual in Section 9.1. Here, we're looking at the choice of the 5 brands, and thus create new "alternatives", labelled 11, 12, .., 23

Stephane

Re: Hybrid model with OL

Posted: 06 Apr 2022, 01:36
by Yashin
Thank you Sir