High Dimensional Generalized Fermat Point Solution Based on PGSA

The optimal aggregation point of two-dimensional data (generalized Fermat points) is extended to the high-dimensional space, and a partitioning clustering algorithm, k-Fermat clustering, which takes the generalized Fermat points as the center point of the optimal clustering is proposed. Firstly, the uniqueness of Fermat points as clustering center points is analyzed theoretically. Secondly, the simulated plant growth algorithm (PGSA) is established to solve Fermat points of high-dimensional data and guide the clustering algorithm system. Finally, the complexity of the algorithm is strictly proved. In this paper, a new method for unsupervised learning is proposed to improve the theory that there is no optimal cluster center point in current partitioning clustering algorithms.