Develop Accurate and Credible Supported Value

Now, an analyst can support quantitative adjustments, deriving them from Elastic Net Regression (ENR) coefficients, estimated with a minimum of four observations and an unlimited number of predictor variables. Alternatively, an analyst may submit the ENR output, which contains predicted value, to support a value conclusion.

The Elastic Net Regression (ENR) program handles situations where the number of predictor variables is equal to or exceeds the number of observations.

The Ordinary Least Squares (OLS) regression program is also available for use when the analyst has a data set with many more observations than predictor variables.

References: Zou, H. and Hastie, T. (2005) Regularization and Variable Selection via the Elastic Net. Journal of the Royal Statistical Society, Series B, 67, Part 2, 301-320. Alboukadel, K. (2017) Penalized Regression: Ridge, Lasso, and Elastic Net. Machine Learning Essentials: Practical Guide in R, ed 1, Lexington, KY: STHDA. 87-95.