Dr. Kenneth M. Lusht, Real Estate Valuation: Principles and Applications (State College, PA: KML Publishing, 2001), 119–138, identified as the three techniques of the sales comparison approach: regression analysis, direct sales comparison, and statistical analysis. Using the Elastic Net Regression (ENR) program, regression analysis and the traditional direct sales comparison could be used to develop an opinion of a defined standard of value for real property. In applying the traditional direct sales comparison technique, adjustments for differences in property productivity features are derived using the coefficients of the estimated regression model. In fact, given that the ENR handles small sample that we typically encounter in the traditional direct sales comparison analysis, there would be no need to also apply the traditional direct sales comparison.
Dr. A. Ason Okoruwa, in a working paper titled, “Elastic Net Regression, a Penalized Ordinary Least Squares Regression: Applicable when Number of Predictor Variables Exceeds Number of Observations,” applied the penalized ordinary least squares regression technique in the valuation of a multi-family property and a single-family property. Three models each, using four observations, five observations, and six observations, were estimated for the multi-family property and single-family property, respectively. A link to the paper follows:
Dr. A. Ason Okoruwa, in a paper titled, “How to Interpret Regression Coefficients and Calculate Adjustments for Differences in Property Productivity Features,” published in the Winter 2018 edition of The Appraisal Journal, demonstrates how to calculate adjustments from regression coefficients to use in the traditional direct sales comparison. The value estimates from both techniques are close. A link to the article follows: