Development of a speech emotion recognizer for large-scale child-centered audio recordings from a hospital environment.