Machine Learning Algorithms for Automatic Classification of Marmoset Vocalizations.

Hjalmar K Turesson,Sidarta Ribeiro,Danillo R Pereira,Victor Hugo C De Albuquerque,João P Papa,Michael Smotherman

doi:10.1371/journal.pone.0163041

Abstract

Automatic classification of vocalization type could potentially become a useful tool for acoustic the monitoring of captive colonies of highly vocal primates. However, for classification to be useful in practice, a reliable algorithm that can be successfully trained on small datasets is necessary. In this work, we consider seven different classification algorithms with the goal of finding a robust classifier that can be successfully trained on small datasets. We found good classification performance (accuracy > 0.83 and F1-score > 0.84) using the Optimum Path Forest classifier. Dataset and algorithms are made publicly available.

Highlights

The common marmoset (Callithrix jacchus) is a species of arboreal New World monkeys native to the northeast region of Brazil
Some of the reasons for this popularity are that marmosets have a similar disease susceptibility profile to humans, are relatively easy to handle, have a high reproductive rate, and that important genetic and neuroscience research tools already exist [7]
In order to find a robust method for the automatic classification of marmoset vocalizations, we compared the classification performance of seven different algorithms

Summary

Introduction

The common marmoset (Callithrix jacchus) is a species of arboreal New World monkeys native to the northeast region of Brazil. This species is becoming an increasingly important primate model of a number of human diseases, as well as for basic research in neuroscience [1, 2] and genetics [3]. Some of the reasons for this popularity are that marmosets have a similar disease susceptibility profile to humans, are relatively easy to handle, have a high reproductive rate, and that important genetic and neuroscience research tools already exist [7]. Given the widespread use of marmosets in laboratories, methods to reliably monitor health and behavior in captive colonies are of a high priority

Methods

Results

Conclusion