A multimodal approach to chaotic renewable energy prediction using meteorological and historical information