Applications of Ridge Regression in Forestry

Author: Bare B. Bruce   Hann David W.  

Publisher: Society of American Foresters

ISSN: 0015-749X

Source: Forest Science, Vol.27, Iss.2, 1981-06, pp. : 339-348

Disclaimer: Any content in publications that violate the sovereignty, the constitution or regulations of the PRC is not accepted or approved by CNPIEC.

Previous Menu Next

Abstract

Describes the use of ridge regression for dealing with multicollinearity in multiple linear regression. Ridge regression is reviewed and three criteria for selecting the "best" ridge estimator--ridge trace, variance inflation factor, and determinant of the correlation matrix--are discussed. The first application demonstrates the use of ridge regression for selecting independent variables during the development of a ponderosa pine basal area growth model. This use of ridge regression produced a meaningful predictive model with interpretable coefficients. The second application uses ridge regression to develop a descriptive model for estimating bare land values in the Douglas-fir region. The objective was to produce precise and stable estimates of model parameters and not to predict the dependent variable. The resulting bare land value estimates fall in the range of values produced by other techniques. Forest Sci. 27:339-348.