PREDICTION OF TREE STEMS TO IMPROVE EFFICIENCY IN AUTOMATIZED HARVESTING OF FORESTS

The problem of predicting future observations on a statistical unit gi ven past measurements on the same and other similar units is frequentl y encountered in practical applications. When computer-based marking f or bucking routines is used in a forest processor, it is usually not f easible to run the whole tree stem through the measuring device before the first cutting decisions have to be made. However, for optimal con version of single stems into smaller logs, the whole stem should be me asured in advance. To this end we have developed a predictor for the u nknown part of the stem under the mixed model for repeated measurement s. Our prediction-based approach provides an eminently satisfactory so lution to this important marking for bucking problem under incomplete information.