Tk-merge: Computationally Efficient Robust Clustering Under General Assumptions
This provides a computationally efficient solution for robust clustering in various domains, but it appears incremental as it builds on existing trimmed k-means and hierarchical methods.
The authors tackled general-shaped clustering under weak assumptions with a two-step hybrid robust algorithm, achieving low computational complexity and effective cluster identification even with data contamination, and it outperformed state-of-the-art methods in simulations and real-world applications like image analysis and GPS data.
We address general-shaped clustering problems under very weak parametric assumptions with a two-step hybrid robust clustering algorithm based on trimmed k-means and hierarchical agglomeration. The algorithm has low computational complexity and effectively identifies the clusters also in presence of data contamination. We also present natural generalizations of the approach as well as an adaptive procedure to estimate the amount of contamination in a data-driven fashion. Our proposal outperforms state-of-the-art robust, model-based methods in our numerical simulations and real-world applications related to color quantization for image analysis, human mobility patterns based on GPS data, biomedical images of diabetic retinopathy, and functional data across weather stations.