| For this study, a classification program capable of identifying four hydrophilid beetle species from vocalizations in under water hydrophone recordings was created. Within single-species recordings, classification accuracy ranged from 81-98%. Mathematical features, based on the frequency content of exemplar beetle vocalizations, were used to compare hydrophilid vocalizations with new sound data in Matlab™ and classify sounds as a beetle species, beetle distress call, or noise. |
