00:01
So when we have a residual plot here and we need to check if everything is okay, basically the residual will give us the difference between the observed value that we have in our data with the predicted value that we got using the regression, right? so if everything is okay, the residues will be centered around zero and will like be here at random.
00:27
We don't see any pattern here, it's random and it is close to zero.
00:35
And basically in terms of linearity, we should have at least this kind of randomness here.
00:44
We can still have a large values here, but this is the key to say that we have a linear relationship between the variable.
00:53
If you don't have this, you can say that you have a problem with your model and you should try to using, i don't know, adding more variables to your model or using some transformations, that kind of stuff.
01:07
If for some reason your residual is like going a straight line or going a curve, this means that your model is not okay.
01:18
So considering here the options that we have, which is that if a residual plot has residues that appear to be random, then it's okay.
01:25
It is okay to assume that we have a linear relationship between the predictor and the response variable, which is right...