Estimating the Characteristics of a Marked Stand Using k-Nearest- Neighbour Regression
Authors
Mikko Tommola
Joensuu, Finland
Mika Tynkknyen
Joensuu, Finland
Jussi Lemmetty
Joensuu, Finland
Pertti Harstela
Joensuu, Finland
Lauri Sikanen
Joensuu, Finland
Abstract
The purpose of this study was to develop the k-nearest-neighbour method as a wood procurement planning tool. Traditionally, sampling measurement of standing trees has been used to obtain advance information on marked stands. In this study, key figures such as sawtimber/pulpwood ratio in pine and spruce stands, diameter and height distribution in spruce stands, diameter and quality distribution in pine stands, and quality distribution by diameter classes in pine stands were estimated using k-nearest-neighbour regression. The material consisted of 716 stands. Stands were located in the eastern Finland. Information regarding every stand was collected from the information system of one large Finnish timber-procurement organization. The accuracy of the k-nearest-neighbour method was compared with the traditional planning inventory method and stand inventory method. The created model was found to be a useful tool in the planning of wood procurement.