The Use of Mathematical and Statistical Analysis in Selecting of Best Regression Equation for Quercus Infectoria G.Olivier

Abstract

The purpose of this study is to use mathematical and statistical analysis in testing the performance ability of standard volume models in volume prediction of Gall Oak trees. A sample of 205 non - destructive trees was purposely selected from Duhok Governorate, Kurdistan region, Iraq. Of which 157 were used for estimating the parameters of regression models and the rest (48) trees served as validation data. Using computer facilities, the available software packages, and tools including STATGRAPHIC Centurion and Excel tool, the parameters of 28 models were estimated. A multistage procedure was used for screening the generated regression equations and selecting the most appropriate one, which underwent validation with independent data. In this study, a new criterion was proposed and used for testing the accuracy of regression equations in the prediction of the dependent variable. At the end of this process, the regression equation Vt = 0.00442+0.0108D^1.917H was finally selected as the most appropriate one that fits the data.