Corpus-based generation of head and eyebrow motion for an embodied conversational agent