Application of a Modified Statistical Segmentation Method to Timber Machine Strength Grading

Authors

  • Frédéric Rouger

Keywords:

Timber, machine grading, statistics, classification

Abstract

In this paper, a new method for deriving grading rules is given. This method is based on the multiple regression and discrimination techniques by binary prediction trees, which are of high interest for classification purposes. A modification to the existing practice that tends to predict the bending modulus of rupture (MOR) is presented. This method consists in creating a nominal variable, the "optimal strength class value," which contains the information necessary to enter into a strength class system, i.e., the density, the bending modulus of elasticity (MOE), and the MOR. A comparison between a regression technique aimed at predicting the MOR and a discrimination technique aimed at predicting the optimal strength class assignment illustrates the innovative aspect of this method.

References

Assoc. Française de Normalisation (AFNOR). 1996. NFB 52001.4: Regles d'utilisation du bois dans la construction. Partie 4: Classement visuel pour l'emploi en structure des principales essences resineuses et feuillues. Version 1996 privisoire et non publiée. AFNOR, Paris, France.nAssoc. Française de Normalisation (AFNOR). 1995. NF EN338: Bois de structure—classe de resistance. P21353. (Structural timber—Strength classes).nBreiman, L, J. H. Friedman, R. A. Olshen, and C. J. Stone. 1984. Classification and regression trees. Wadsworth Int. Group.nFewell, A. R. Undated. Derivation of setting for stress grading machines. A brief description of the approach used to derive settings in the U.K. Undated Report. Building Research Establishment, Watford, UK.nMorgan, J. N., and A. Sonquist. 1963. Problems in the analysis of survey data and a proposal. J. Am. Statist. Assoc. 58:415-435.nRouger, F., C. De LaFond, and A. El. Ouadrani. 1993. Structural properties of French grown timber according to various grading methods. 26th CIBW18 Meeting, August 1993, Athens, GA.n

Downloads

Published

2007-06-19

Issue

Section

Research Contributions