A Study on Robust Hierarchical Clustering Using LOF

Abstract

Hierarchical clustering is a method of cluster partitioning by successively merging clusters based on dissimilarity. In clustering, outliers can significantly impact the quality of the resulting cluster partition. This paper proposes a hierarchical clustering based on weighted dissimilarity using the Local Outlier Factor (LOF) to improve robustness against outliers. The proposed method mitigates the effect of outliers in the clustering process by incorporating LOF-based weighting into the dissimilarity calculation. Experiments on artificial and real datasets were conducted to demonstrate the effectiveness of the proposed method. The experimental results show that the proposed method improves the robustness of clustering compared to conventional hierarchical clustering methods.