Algorithms for Artificial Neural Networks with Interval Target Tuning for the Task of Classifying Cammeo and Osmancik Rice Varieties

This paper investigates the application of intelligent learning algorithms for artificial neural networks (ANN) with interval target tuning for the task of classifying Cammeo and Osmancik rice varieties. The data used includes 3810 rice grain samples for which seven morphological features were measured. The core concept of the method is the creation of a multi-level hierarchical network consisting of various typical ANN modules interacting in a hierarchical structure based on a specially developed algorithm. The process of network synthesis and training is conducted, and the effectiveness of the proposed approach is analyzed. The results show that this method allows achieving high classification accuracy.