Comparing nearest neighbor configurations in the prediction of species-specific diameter distributions

摘要: We examine how the configurations in nearest neighbor imputation affect performance of predicted species-specific diameter distributions. The simultaneous for all tree species and separate by are evaluated with total volume calibration as a prediction method distributions. This study considers predictions distributions Finnish boreal forests means airborne laser scanning (ALS) data aerial images. aim was to investigate different non-parametric (NN) determine changes error rates timber assortment volumes indices Non-parametric NN used modeling applied two ways: (1) were at same time imputation, (2) one imputation. Calibration regression-based both cases. results indicated that significant can be achieved selection responses, volume, species. Overall, response variables improved distribution rates. most successful