IPez: An expert system for the taxonomic identification of fishes based on machine learning techniques

摘要

The taxonomic identification of fishes is a time-consuming process for those who are not specialists and, therefore, are more liable to make mistakes. We measured morphometric characters in more than 8900 individuals belonging to 6 classes, 43 orders, 192 families, 510 genera and 847 marine and freshwater species. The aim was to determine if the taxonomic identification of juvenile and adult fishes is possible using these measurements. We developed the expert system IPez, which is based on machine learning techniques, and found that, when the number of individuals measured of a species and, hence, included in the database of IPez, is higher than approximately 15 individuals, IPez identifies correctly 100% of new individuals of this species that are not included into the database. Moreover, besides helping in the taxonomic identification of fish, this software allows the determining of the main morphometric characters that have promoted or are promoting divergence among closely related species. The software is free and available at the web page http://www.ipez.es/index%20ingles.html. We suggest increased international collaboration to introduce more species into IPez.