A Unified BERT-CNN-BiLSTM Framework for Simultaneous Headline Classification and Sentiment Analysis of Bangla News
The research proposes a hybrid BERT-CNN-BiLSTM model to simultaneously classify Bangla news headlines and determine their sentiment. This unified framework combines contextual word embeddings from BERT with local feature extraction via CNN and long-range dependency modeling through BiLSTM, achieving state-of-the-art results for low-resource language classification.