An Empirical Evaluation of Two Novel Linkage Criteria for Hierarchical Agglomerative Clustering
Leonardo Ramos Emmendörfer · 2019
Hierarchical agglomerative clustering (HAC) is among the most widely adopted algorithms in unsupervised learning. The choice of linkage criteria highly influences the results of HAC. This paper presents and evaluates two novel linkage criteria for HAC. The evaluation is performed on 12 datasets from the literature. Novel criteria are compared to three reference methods from the literature: single linkage, complete linkage, and average linkage.