The topic of this thesis is utilization of three types of modern regression methods (linear mixed effects model, generalized additive model and classification and regression trees) to model height-diameter relationship. Results revealed that all three regression methods can be, under certain conditions, applicable as an alternative approach for height-diameter relationship modeling. The mixed effects model performed best, when random effects of the intercept and regression coefficient were taken into account along with the age as a second order predictor (or its natural logarithm). The generalized additive model performed best when LOESS smoothing function was applied. The regression tree method prediction was optimal when diameter at breast height, mean height, age of a forest stand and site index were included to the model as predictors.
| Identifer | oai:union.ndltd.org:nusl.cz/oai:invenio.nusl.cz:249349 |
| Date | January 2014 |
| Creators | Adamec, Z. |
| Source Sets | Czech ETDs |
| Language | Czech |
| Detected Language | English |
| Type | info:eu-repo/semantics/doctoralThesis |
| Rights | info:eu-repo/semantics/restrictedAccess |
Page generated in 0.0131 seconds