An Effective Clustering Algorithm Using Adaptive Neighborhood and Border Peeling Method

Traditional clustering methods often cannot avoid the problem of selecting neighborhood parameters and the number of clusters, and the optimal selection of these parameters varies among different shapes of data, which requires prior knowledge. To address the above parameter selection problem, we propose an effective clustering algorithm based on adaptive neighborhood, which can obtain satisfactory clustering results without setting the neighborhood parameters and the number of clusters. The core idea of the algorithm is to first iterate adaptively to a logarithmic stable state and obtain neighborhood information according to the distribution characteristics of the dataset, and then mark and peel the boundary points according to this neighborhood information, and finally cluster the data clusters with the core points as the centers. We have conducted extensive comparative experiments on datasets of different sizes and different distributions and achieved satisfactory experimental results