The following are the assumptions:
-
The population is represented by the sample data used to fit the model.
-
X and the mean of Y have a linear connection.
-
For every value of X, the residual variance is the same (homoscedasticity).
-
Observations are unrelated to one another.
-
Y is regularly distributed for any value of X.
The results will become redundant if these assumptions are violated to the extreme. Small deviations from these assumptions will increase the estimate’s bias or variance.