Modern Methods of Intelligent Multidimensional Data Analysis in Smart Agro Tasks: From Classical Clustering to Hierarchical MST-Taxonomy of Crop Stress Tolerance

The article presents an analytical review of the development of graph-theoretic algorithms and their application in digital agriculture (Smart Agriculture). The evolution of methods from classical clustering (K-Means) to algorithms for the decomposition of Minimum Spanning Trees (MST) is considered. Based on the analysis of global research, a scientific gap in existing approaches related to the static nature of threshold graph division criteria is substantiated. The concept of Eshankulov's hybrid algorithm is proposed, which utilizes a dynamic recursive recalculation of local subtree metrics to construct a multi-class taxonomy of crop stress tolerance under soil and environmental stress conditions.