Abstract: This research examines the effectiveness of hybrid deep learning architectures for sentiment classification in Thai-language social media, focusing on applications in crisis management. The study proposes and assesses eight configurations of a Hybrid Bidirectional Gated Recurrent Unit (HBiGRU) architecture, employing a specialized GRU-Conv1D-BiGRU pipeline.
Sasiphan Nitayaprapha (2026);
Investigating Hybrid GRU-Conv1D-BiGRU Architectures for Enhanced Sentiment Classification in Thai Crisis News Tweets ;
International Journal of Scientific and Research Publications (IJSRP)
16(5) (ISSN: 2250-3153),
DOI: http://dx.doi.org/10.29322/IJSRP.16.05.2026.p17314
