DendroLikeness: an R package to compare dendrograms
Artículo
Te invitamos a leer el artículo "DendroLikeness: an R package to compare dendrograms" publicado en Zenodo., a cargo del profesor investigador de la Unidad de Genómica Avanzada del Cinvestav Dr. Octavio Martínez de la Vega, Profesor Investigador de la UGA.
Autores: Octavio Martínez de la Vega
-
Laboratorio de Biología Computacional de la Unidad de Genómica Avanzada (UGA) del Cinvestav, Mexico
Felicitamos al estudiantado y profesorado que contribuyeron en esta investigación por su arduo trabajo.
Summary:
If you use R to produce tree diagrams (dendrograms) you could find this package useful. Dendrograms are the result of applying hierarchical clustering methods to data.
When, for a single data set, you use the R functions "dist()" and then "hclust()" with different methods, you could obtain very different dendrograms. Comparing different dendrograms just by looking at their plots is complicated when the number of units (entities) is large. The R package DendroLikeness gives you an easy solution to that problem by having functions to obtain the topology of each dendrogram (this is, the clusters that exist in each graph) and then detecting the clusters shared by two (or more) dendrograms. You will obtain different measures of likeness which will make easier for you to see the differences between dendrograms and thus decide which distance and clustering algorithm is better to show in your scientific publication.