Skip to main content
. Author manuscript; available in PMC: 2022 Jun 1.
Published in final edited form as: J Biomed Inform. 2021 Apr 20;118:103788. doi: 10.1016/j.jbi.2021.103788

Table 4. Algorithm-distance pairs implementing single distance metrics with 3 clustering algorithms on single data-type simulations.

Each distance metric is associated with optimal suitability for certain data types. Some distance metrics can be implemented on multiple data types, while others have type-restricted implementation. Distance metrics were applied to all permitted data types.

Algorithm Distance Data Type Suitability Data Type Implementation
Agglomerative hierarchical clustering with Ward’s method Jaccard Binary (asymmetric) Binary
Sokal-Michener Binary (symmetric) Binary
Gower Nominal;
categorical		 Categorical
Manhattan Ordinal;
continuous;
binary		 Binary, categorical, continuous
Euclidean Continuous;
binary		 Binary, categorical, continuous
Partitioning Around Medoids (PAM) (k-medoids) Jaccard Binary (asymmetric) Binary
Sokal-Michener Binary (symmetric) Binary
Gower Nominal;
categorical		 Categorical
Manhattan Ordinal;
continuous;
binary		 Binary, categorical, continuous
Euclidean Continuous;
binary		 Binary, categorical, continuous
Self-organizing maps Tanimoto Binary Binary
Manhattan Ordinal;
continuous;
binary		 Binary, categorical, continuous
Euclidean Continuous;
binary		 Binary, categorical, continuous