Natural Hierarchical Cluster Analysis by Nearest Neighbors with Near-Linear Time Complexity
This provides a more efficient and universal hierarchical clustering method for data analysis, though it appears incremental as it builds on nearest neighbor approaches.
The authors tackled the problem of hierarchical clustering by proposing a nearest neighbor-based algorithm that defines clusters purely from the dataset, independent of iterative processes, and demonstrated near-linear time and space complexity for certain datasets.
We propose a nearest neighbor based clustering algorithm that results in a naturally defined hierarchy of clusters. In contrast to the agglomerative and divisive hierarchical clustering algorithms, our approach is not dependent on the iterative working of the algorithm, in the sense that the partitions of the hierarchical clusters are purely defined in accordance with the input dataset. Our method is a universal hierarchical clustering approach since it can be implemented as bottom up or top down versions, both of which result in the same clustering. We show that for certain types of datasets, our algorithm has near-linear time and space complexity.