Annika Kangas , Kari T. Korhonen

Generalizing sample tree information with semiparametric and parametric models.

Kangas A., Korhonen K. T. (1995). Generalizing sample tree information with semiparametric and parametric models. Silva Fennica vol. 29 no. 2 article id 5553. https://doi.org/10.14214/sf.a9204

Abstract

Semiparametric models, ordinary regression models and mixed models were compared for modelling stem volume in National Forest Inventory data. MSE was lowest for the mixed model. Examination of spatial distribution of residuals showed that spatial correlation of residuals is lower for semiparametric and mixed models than for parametric models with fixed regressors. Mixed models and semiparametric models can both be used for describing the effect of geographic location on stem form.

Keywords
forest inventories; mixed models; volume; nonparametric models; semiparametric models

Published in 1995

Views 2985

Available at https://doi.org/10.14214/sf.a9204 | Download PDF