in regression , you can get it by finding the feature importance function . highest value(beta) has highest importance. You betas here could be brand , quality , service etc ;
The calculation is based directly on interpretation of regression coefficients. I believe Sawtooth Software has some user manual type of documents on this. If you need a more academic take, Allenby et al (2014) is a good reference. Regular practitioners may have problems following the paper.
Hi again Sir , I am not a specialist at all BUT I think COINJOINT ANALYSIS is a tool that allow us to know which attribute does the customer give more value and thus retailers will product more products having that or those attribute(s) Also , would like to know how to select those attributes to TEST from the begining to be realistic because if the attributes are not realistic then the customers answers will not be realistic.. Also , for the baseline , I guess you take the MINIMUM of each quantitive variable but what about qualitative variable what if you didn't know that AIRBNB is lower valued than Hilton and Mariott (what will you do in this case ?) Also , for me this more a COMPETITVE ANALYSIS as you are setting AIRBNB as baseline. If you have something instead AIRBNB then I guess we will not have the same feature importance score . Same thing if we change one more time the baseline . So the question for the importance feature score will be stable after selecting several hotels and take the mean of EACH feature importance scores realized with each baseline. What do you think ? Thank you in advance for the answers
But how is the regression done exactly? How are the users choices transformed into predicting and predicted variables? How do you define the logit function? You skipped over that part :(
This video deserves more views and your channel deserves much more subscribers.
Thank you so much
Wow, thank you!
This was really helpful, thank you very much ❤
Thank you, it has been very helpful, just one question How do you obtain the IMPORTANCE COEFFICIENTS???
in regression , you can get it by finding the feature importance function . highest value(beta) has highest importance. You betas here could be brand , quality , service etc ;
Muchas gracias 🙏🏻🙏🏻🙏🏻
Sir , Thank you sooo much , really well explained and useful video. Liked and Subscribed.
Do you have a source for the calculation of the willingness to pay in the last part of your video?
The calculation is based directly on interpretation of regression coefficients. I believe Sawtooth Software has some user manual type of documents on this. If you need a more academic take, Allenby et al (2014) is a good reference. Regular practitioners may have problems following the paper.
@@YeHu4ps Hi, is true that this algebraic approach lead to overstate the WTP?
Thank u so much for this video. It is very helpful and clear
Hi again Sir , I am not a specialist at all BUT I think COINJOINT ANALYSIS is a tool that allow us to know which attribute does the customer give more value and thus retailers will product more products having that or those attribute(s)
Also , would like to know how to select those attributes to TEST from the begining to be realistic because if the attributes are not realistic then the customers answers will not be realistic..
Also , for the baseline , I guess you take the MINIMUM of each quantitive variable but what about qualitative variable what if you didn't know that AIRBNB is lower valued than Hilton and Mariott (what will you do in this case ?)
Also , for me this more a COMPETITVE ANALYSIS as you are setting AIRBNB as baseline. If you have something instead AIRBNB then I guess we will not have the same feature importance score . Same thing if we change one more time the baseline . So the question for the importance feature score will be stable after selecting several hotels and take the mean of EACH feature importance scores realized with each baseline.
What do you think ?
Thank you in advance for the answers
But how is the regression done exactly? How are the users choices transformed into predicting and predicted variables? How do you define the logit function? You skipped over that part :(
thank you very much