The neural network classification of false killer whale (Pseudorca crassidens) vocalizations

This study reports the use of unsupervised, self-organizing neural network to categorize the repertoire of false killer whale vocalizations. Self-orga nizing networks are capable of-detecting patterns in their input and partit ioning those patterns into categories without requiring that the number or types of categories be predefined. The inputs for the neural networks were two-dimensional characterization of false killer whale vocalizations, where each vocalization was characterized by a sequence of short-time measuremen ts of duty cycle and peak frequency. The first neural network used competit ive learning, where units in a competitive layer distributed themselves to recognize frequently presented input vectors. This network resulted in clas ses representing typical patterns in the vocalizations. The second network was a Kohonen feature map which organized the outputs topologically, provid ing a graphical organization of pattern relationships. The networks perform ed well as measured by (1) the average correlation between the input vector s and the weight vectors for each category, and (2) the ability of the netw orks to classify novel vocalizations. The techniques used in this study cou ld easily be applied to other species and facilitate the development of obj ective, comprehensive repertoire models. (C) 1998 Acoustical Society of Ame rica. [S0001-4966(98)03312-8].