The exponential growth of social evaluations of services has led many researchers to focus on emotion analysis. Having so much data helps analyze end-user behavior to improve QoS. Text categorization is a prominent language processing research area that organizes the disorganized text into useable categories. The LSTM and CNN models are widely employed in text-based categorization applications and produce reliable results. By contrast, LSTM-based models gain long-term connections between text sequences and so are better suitable for text classification. In this case, the hybrid approach may memorize classification, slowing down the training process. This work proposes an optimal attention-focused BiLSTM and ConvNet model. The suggested model is trained on two independent datasets to validate its performance. The proposed attention-based model outperformed previous deep learning algorithms. Compared to existing machine learning methods, the proposed approach is more accurate.
You may also start an advanced similarity search for this article.