Recognizing emotions in spoken dialogue with hierarchically fused acoustic and lexical features | IEEE Conference Publication | IEEE Xplore