Machine Learning Algorithms Can Predict Emotional Valence Across Ungulate Vocalizations

This folder contains extracted individual animal calls ('SoundsDatabase.zip'), labelled according to species, context of production, and emotional valence (i.e., negative or positive). The repository contains a total of 3,181 contact calls. The document 'SoundsRepartition.xlsx' includes detailed information on each extracted call, such as the animal ID, species, call type, emotional valence, context of production, number of calls, and references to the studies from which these calls were collected. Additionally, 'SoundsDataBaseKey.xlsx' contains a database of the extracted acoustic parameters, including duration, frequency, amplitude modulation, spectral centroid, and more.