Question: Using some computer program, such as the Minitab macro RIDGE, construct the ridge trace for the last 26 observations in Table and deter-mine if the ridge constant seems reasonable relative to the trace. If not, explain what probably causes this ridge constant k = p^σ2/ß^'ß to be inappropriate for this (partial) data set. Should all six of the variables be used? If some variables can be deleted, should ridge regression be applied to the subset that remains? If so, construct the ridge trace and compute the HKB ridge constant.
