
LINEAR MODEL 69
2.2.3 Regression Diagnostics
Calling lm() always returns a result if it is appropriate for the data, but it will also return
a linear result if the linear model is not adequate. We need additional diagnostics to tell us
whether the model is reliable and usable.
Exercise 2.3
Let
yy <- 2.5*x +0.01 x^2 + err
What are the results you get if you do a regression using the (incorrect)
regression model yy ∼x ? Do you get any hints that this model is not
adequate?
The function lm() not only gives an estimation for the linear model, but also provides a
series of diagnostics that can help to judge whether the model assumptions are acceptable. A
represen ...