A Simple Non-Linear Weighted Regression Computer Program for Forest BiomassEstimations

1988

3 Abb., 16 Lit. Ang.

Unselbständiges Werk

2819

200014576

The paper presents a computer program destined to fit experimental data through the nonlinear models: (Formel). This kind of model is particularly interesting for forestry biomass estimations (where Y = dry weight; X1 = diameter; X2 = height of the tree). The program takes into account: The research of optimal values for both exponents; the possibility to weight the residuals of the regression with a power function of the explanatory data (Formel),as the case may be; the calculation of the confidence bounds (level 95 p. 100) for estimated values; the calculation of the error when a such an equation is used to predict an estimation of the total biomass of a population. The practical interest of the program stands in its flexibility (conversational program with possibilities of choice and a partical presentation of the important intermediate results on the screen - helping to choose the best option) and in the fact that it is a self sufficient program for the PC.