End-to-end linkage (EEL) clustering algorithm: a study on the distribution of Meissner corpuscles in the skin

A novel hierarchical clustering algorithm was applied to the distribution of Meissner corpuscles in the skin of mammals. This method, called end-to-end linkage (EEL), is useful for grouping data that consists of chain-like contingencies in the multivariable space. Unlike the traditional techniques which uncover hyperspherical clusters (e.g. single linkage), EEL considers the shortest distance between the predefined end pairs of the two clusters as an inter-group distance. This scheme allows characterizing the internal structure of data better than other hierarchical techniques. The anatomical data used in the case study is important for studying the sense of touch. The results show a substantial improvement over the traditional single-linkage method. On average, the number of correctly classified corpuscles is increased to twice the number identified by the single-linkage method. EEL can also be used for analyzing other sensory modalities where geometric relationships need to be explored. In addition, the report contains corpuscle density and epidermal-ridge width data obtained from several species.