Decision Support System for Bat Identification using Random Forest and C5.0

Abstract

Morphometric and morphological bat identification are a conventional method of identification and requires precision, significant experience, and encyclopedic knowledge. Morphological features of a species may sometimes similar to that of another species and this causes several problems for the beginners working with bat taxonomy. The purpose of the study was to implement and conduct the random forest and C5.0 algorithm analysis in order to decide characteristics and carry out identification of bat species. It also aims at developing supporting decision-making system based on the model to find out the characteristics and identification of the bat species. The study showed that C5.0 algorithm prevailed and was selected with the mean score of accuracy of 98.98%, while the mean score of accuracy for the random forest was 97.26%. As many 50 rules were implemented in the DSS to identify common and rare bat species with morphometric and morphological attributes.