| Abstract: | We provide a brief summary of the theory of ridge regression, regress reported rates of congenital syphilis on three stages of adult syphilis using both ridge regression and least squares techniques, and discuss the results. We also develop a principal components regression model for the same example and compare the results. The example demonstrates the usefulness of ridge regression to handle multicollinearity in data. Problems of overestimation and instability associated with ordinary least squares estimates of these non-orthogonal data diminish with use of the ridge regression techniques. |
