An Arabic language benchmark and hybrid BERT–BiGRU–CapsNet model for optimism and pessimism detection

Tawfeeq Alsanoosy et al.