CNMBI: Determining the Number of Clusters Using Center Pairwise Matching and Boundary Filtering

arXiv cs.CV / 3/31/2026

📰 NewsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper introduces CNMBI, a method for determining the number of clusters without assuming prior distribution information about the data.
CNMBI models cluster-center comparison through a dynamic positional-behavior process and uses bipartite graph theory to make the matching efficient without requiring full clustering outputs.
It incorporates sample-specific confidence to actively filter out low-confidence samples, which the authors claim has not been addressed in prior cluster-number determination approaches.
Experiments on complex datasets including CIFAR-10 and STL-10 show CNMBI is robust and more flexible with respect to data dimensionality and cluster shape.
Comparative studies against state-of-the-art competitors indicate CNMBI achieves superior performance across multiple challenging benchmarks.

Abstract

One of the main challenges in data mining is choosing the optimal number of clusters without prior information. Notably, existing methods are usually in the philosophy of cluster validation and hence have underlying assumptions on data distribution, which prevents their application to complex data such as large-scale images and high-dimensional data from the real world. In this regard, we propose an approach named CNMBI. Leveraging the distribution information inherent in the data space, we map the target task as a dynamic comparison process between cluster centers regarding positional behavior, without relying on the complete clustering results and designing the complex validity index as before. Bipartite graph theory is then employed to efficiently model this process. Additionally, we find that different samples have different confidence levels and thereby actively remove low-confidence ones, which is, for the first time to our knowledge, considered in cluster number determination. CNMBI is robust and allows for more flexibility in the dimension and shape of the target data (e.g., CIFAR-10 and STL-10). Extensive comparison studies with state-of-the-art competitors on various challenging datasets demonstrate the superiority of our method.