Multicollinearity Remedial Techniques in Model-Building

The dire need for renewable Green-Energy resources, the vitality for sustainable forest, its management and practices, just to name a few, are examples of the importance for an estimation tool. Hence, a modeling approach is developed, based on selected biomass equations adopted in Forest Science. Models using the multiple regression techniques are employed. Correlation Coefficients of variables are found to have multicollinearity effects. Illustrations on the algorithm of the remedial techniques are exemplified which focuses on the removal of absolute coefficients values of more than 0.95. Significant variables with their possible interactions are selected using statistical tests. Best model is selected based on the eight selection criteria (8SC). The best regression model without multicollinearity is found to give a better estimation with different major contributions from mensuration data. Keywords: Biomass Equation; Correlation Coefficients; Multiple Regression; Multicollinearity; Best Regression Model. 2010 Mathematics Subject Classification: 62J02