You need to have all possible combinations. If there are too much variables and you df is getting low you can replace the variables with the squared predicted values (e2= ꞵ0 + ꞵ1 ŷ2) ; this is called the reduced white test.
Hi Muhammad Indeed, my students get the explication in my course. This video's are a brief overview from what I teach. I can create them much faster without voice, but maybe I should experiment with voice ;-) Because I teach in Dutch, the video's are in this language.
You need to have all possible combinations. If there are too much variables and your df is getting low you can replace the variables with the squared predicted values (e2= ꞵ0 + ꞵ1 ŷ2) ; this is called the reduced white test.
what if i have 3 variabels?
axbxc or
axb axc bxc
You need to have all possible combinations. If there are too much variables and you df is getting low you can replace the variables with the squared predicted values (e2= ꞵ0 + ꞵ1 ŷ2) ; this is called the reduced white test.
sorry Ed van Stee there is no voice with your vedio of Heteroskedasticiteit White test
Hi Muhammad
Indeed, my students get the explication in my course. This video's are a brief overview from what I teach. I can create them much faster without voice, but maybe I should experiment with voice ;-)
Because I teach in Dutch, the video's are in this language.
What if I have 5 variabels ?
You need to have all possible combinations. If there are too much variables and your df is getting low you can replace the variables with the squared predicted values (e2= ꞵ0 + ꞵ1 ŷ2) ; this is called the reduced white test.