find residue of functionb good morning friends
Answers
Answer:
A residual is the amount, positive or negative, that the observation differs from the prediction of a regression line.
Once the regression is run, chart the residuals. Examine the plot to see if certain conditions exist. I will give examples there are more. Example
1: the residuals form a “smiley face” (or frown) pattern. This means the variable has a second order effect. By putting in a squared term you capture that variability and improve your model.
Example 2: The residuals form a wedge. There is some effect not being modeled. In this case there are a few different possibilities. One possibility is interaction between variables, there are at least 2 more. No matter what it is a wedge means problems with the model. It may not be fatal but it will be more work.
The main point is the residuals tell you if there is an effect causing variation that the model does not represent. If you can capture more variation that increases R squared, increases the F statistic thus decreasing p-value and will likely make individual regression lines fit better. You would like to see a random distribution of residuals with absolutely no pattern
May be you can get some information from this picture
...
