Machine Learning Algorithms Can Be Trained To Interpret Emotional Valence Across Ungulate Vocalizations Database

Dataset

Description

This folder contains extracted, individual animal calls labeled according to the species, context of production, emotional valence (i.e., negative or positive) and call type when relevant. These calls were collected as part of previous studies listed in the "References" section of the current page. The repository contains a total of 3253 calls. The settings used to extract the acoustic parameters from the audio files and the resulting data frame with extracted parameters are included. The number of calls available per species and context, and accompanying references have also been provided.
Date made available2023
PublisherZenodo

Cite this