Compensating for the loss of future tree values in the model of Fuzzy knowledge units

Abstract

When working with any application domain, it is necessary to grasp and represent the knowledge from this application domain into a suitable form. There is naturally a significant difference between the knowledge gained from natural knowledge, estimation and experience, and the knowledge gained through exact measurement, but it is often necessary to use estimates based on experience in decision-making processes. This is especially important if this representation is to be used in decision support systems using e.g. artificial intelligence (AI) and machine learning (ML). In this paper, we therefore describe a model for valuing solitary trees that allows the use of vague evaluation of input parameters for the evaluation of trees based on fuzzy knowledge units. The creation of the model is based on the parametric method of the Nature Conservation Agency (NCA) and other methods such as CAVAT or FEM, from which the knowledge text is separated. Fuzzy knowledge units (FKU) or Knowledge units (KU) are created from this knowledge text. These FKU are trained according to data from the NCA method and optimized using the MATLAB Tune fuzzy inference system (TUNEFIS). The Adaptive neuro fuzzy inference system (ANFIS) was chosen as the best FKU model. These fuzzy knowledge units are arranged in a hierarchical model of valuing solitary trees, which is implemented in Simulink. The experimental study clearly shows that the proposed model is more detailed in some parameters than a crisp tree evaluation calculator or CAVAT calculator in excel and provides more precise results.