11.67 Before a particular job is accepted, a computer at a major university estimates the cost of running the job to see whether the user’s account contains enough money to cover the cost. As part of the job submission, the user must specify estimated values for two variables: central processing unit (CPU) time and number of lines printed. Although the CPU time required and the number of lines printed do not account for the complete cost of the run, it is though that knowledge of their values should allow a good prediction of job cost. The following model is proposed to explain the relationship of CPU time and lines printed to job cost:

E(Y ) = b0 + b1x1 + b2x2 + b3x1x2 where Y job cost (in dollars)
x1 number of lines printed x2 CPU time Records from 20 previous runs were used to fit this model. The SAS printout is shown in Figure 11.62 a Identify the least-squares model that was fit to the data.
b What are the values of SSE and (estimate of ) for the data?
c Explain what is meant by the statement “This value of SSE [see part (b)] is minimum.”