Artificial Neural Networks: A New Tool for Studying Lemur Vocal Communication

摘要: Previous studies have applied Artificial Neural Networks (ANNs) successfully to bioacoustic problems at different levels of analysis (individual and species identification, vocal repertoire categorization, sound structure) but not nonhuman primates. Here, we report the results applying this tool two important in primate communication. First, apply a supervised ANN classify 222 long grunt vocalizations emitted by five genus Eulemur. Second, use an unsupervised self-organizing network identify discrete categories within black lemurs (Eulemur macaco). Calls were characterized both spectral (fundamental frequency formants) temporal features. The result show only that ANNs are effective for studying also can increase efficiency, objectivity, biological significance classification greatly. advantages over more commonly used statistical techniques applications discussed.