Please enable JavaScript.
Coggle requires JavaScript to display documents.
Multiple Regression (Generally (F = uploaded image (F-test is a “multiple…
Multiple Regression
Generally
-
k - no of predictors
numerator = sum of sq residuals
note that s^2 = standard error
In simp reg, k = 1
Conf Int, but dof = n-k-1
Hypothesis test - same as simple reg, just dof = n-k-1 for p val
-
-
F =
F-test is a “multiple hypothesis test”
that tests the significance of the regression line
ie. Atleast one b significant
Multiple R - coefficient of correlation between the dependent variable and the set of predictors.
Under root of R sq
Adjusted R - Normal r^2 goes up if we add more predictors, this one only goes up if predictors improve model
If different slopes
-
-
If non-linear term P>0.1, no include
Difference
Simple Reg
-
Categorical Variables
-
If has M categories, do M-1 dummy vars
-