Canadian Journal of Forest Research · 2014 · 20 citations · 39 references
Natural ResourcesForestryForest InventoryForest ManagementStatistical ModelingTimber ProductsWood ProductsPoisson RegressionProduct Assortment StrategyStem Diameter VariationStem Volume AssortmentTree Size MetricsZero-inflated DistributionsZero-inflated ModelsZero-inflated Regression
Realistic forestry value chain simulations require accurate representations of each component. For primary processing, this is complicated by the fact that a single raw material is converted into a wide range of lumber products. The aim of this study was to develop statistical models for predicting lumber product assortment from tree size information, while taking into account the high proportion of zeros in the data. Lumber recovery was simulated from a database of 1013 laser-scanned Picea mariana (Mill.) Britton, Sterns & Poggenb. and Abies balsamea (L.) Mill. stems using the sawing simulator Optitek. The number of boards per stem of specific products was modelled with zero-inflated Poisson regression using stem diameter and height as covariates. The number of boards per stem was strongly related to both diameter and height, but also changed according to input prices for lumber products. Zero-inflated models outperformed ordinary Poisson regression in all cases. The developed models will be integrated into simulation tools designed to optimize processes along the entire forest value chain from forest to end user.
39
R: A Language and Environment for Statistical Computing
R Core Team · 2000 · 353K citations · Full text
Natural Sciences, Computer Science, Statistical Computing +10
Likelihood Ratio Tests for Model Selection and Non-Nested Hypotheses
Quang Vuong · Econometrica · 1989 · 6K citations