Analysis of the Modulus of Rupture of Particleboard Subjected to Outdoor Exposure Using a Generalized Linear Mode

Abstract

The Relationship between exposure time and modulus of rupture （MOR） of particleboard subjected to outdoor exposure was examined using simple regression analysis or a generalized linear model （GLM）. As for simple regression analysis, it suggested yields of negative values for predicted mean MOR and lower 95% prediction interval when the exposure time is extended, because of an assumption of linearity and homogeneity of normal distribution. By contrast, for GLM, the predicted mean MOR exponentially decreased, approaching zero. The lower 95% prediction interval did not yield any negative values. The GLM analyzed the relationship between the exposure time and MOR, and the effects of the exposure time on the reduction of MOR were evaluated suitably using the GLM. : outdoor exposure, particleboard, generalized linear model, simple regression analysis, 95% prediction interval.