Labeled frog-call dataset of Yasuní National Park for training Machine Learning algorithms
Labeled dataset of frog calls recorded at Yasuní National Park using directional and omni-directional microphones. We trained a Bayesian classifier with Gaussian Mixture Models and applied it to study the audio in Unidentified Long Recordings. The results of the analysis are in the related article. We showed that the application of a frog-call recognition algorithm allows the estimation of presence-absence of the trained subset of species in real-world recordings of an actual amphibian monitoring made in the wild.
Steps to reproduce
Apply the algorithm described in the related article.