A Comprehensive Comparison of Artificial Neural Networks and Some Regression Modeling Techniques on Dominant Height Predictions of Oriental Spruce in a Mixed Forest, the Northeast Turkey

Background: Dominant height is needed for assessing silvicultural practices in sustainable wood production management. Also, dominant height is used as an important explanatory variable in forest growth and yield models. This paper introduces the evaluation for Artificial Neural Networks and Some Regression Modeling Techniques on Dominant Height Predictions of Oriental Spruce in a Mixed Forest, the Northeast Turkey. Methods: In this study, 873 height-age pairs were obtained from oriental spruce trees in a mixed forest stand. Nonlinear mixed-effects models (NLMEs), autoregressive models (ARM), dummy variable method (DVM), and artificial neural networks (ANNs) were compared to predict dominant height growth. Results: The best predictive model was NLME with single random parameter (root mean square error, RMSE: 0.68 m). The results showed that NLMEs outperformed ARM (RMSE: 1.09 m), DVM in conjunction with ARM (RMSE: 1.09 m), and ANNs (RMSE: from 1.11 to 2.40 m) in majority of the cases. Whereas considering variations among observations by random parameter(s) significantly improved predictions of dominant height, taking into account correlated error terms by autoregressive correlation parameter(s) enhanced slightly the predictions. ANNs generally underperformed compared to NLMEs, ARM, and DVM with ARM. Conclusion: All regression techniques fulfilled the desirable characteristics such as sigmoidal pattern, polymorphism, multiple asymptote, base-age invariance, and inflection point. However, ANNs could not most of these features excluding sigmoidal pattern. Accordingly, ANNs seem to insufficient to assure biological growth assumptions regarding dominant height growth.