Computational Thinking Reflected in K-means Clustering: Taking NetLogo Simulation as an Example

The K-means clustering algorithm, as the core technology of unsupervised learning, is a concentrated embodiment of computational thinking in terms of its inherent logic. This article aims to go beyond the conventional understanding of K-means as a simple data analysis tool, and systematically deconstruct the core elements of decomposition, abstraction, pattern recognition, and algorithm design contained in K-means algorithm from the paradigm of computational thinking. The study used the NetLogo multi-agent simulation platform as the experimental environment, modeling data points and cluster centers as autonomous “agents” to transform abstract algorithm processes into a dynamic and visual computing world. The simulation experiment intuitively revealed that the “allocation update” iteration cycle of the algorithm is essentially a process of parallel execution of simple rules at the subject level, and ultimately the emergence of a globally ordered structure, vividly interpreting the core concept of computational thinking of “building order through rules and iterations”. This article argues that NetLogo simulation is not only a tool for verifying the effectiveness of algorithms, but also a “mirror” that can clearly reflect how computational thinking gradually shapes and solves complex problems. This study ultimately demonstrates that the analysis framework combining K-means algorithm with multi-agent simulation provides a concrete and interactive effective path for understanding and imparting computer thinking, deepens people's understanding of the essence of algorithms, and demonstrates important enlightening value in computer science education methodology.