A novel BiGRUBiLSTM model for Multilevel Sentiment Analysis Using Deep Neural Network with BiGRU-BiLSTM

• Main Authors: Islam, Md Shofiqul, Ngahzaifa, Ab. Ghani
• Format: Conference or Workshop Item
• Language: English
• Published: Springer Singapore 2021
• Subjects: QA75 Electronic computers. Computer science; QA76 Computer software
• Online Access: http://umpir.ump.edu.my/id/eprint/33280/1/A%20novel%20BiGRUBiLSTM%20model%20for%20multilevel%20sentiment%20analysis%20using%20deep%20neural%20network%20with%20BiGRU-%20BiLSTM.pdf; http://umpir.ump.edu.my/id/eprint/33280/; https://doi.org/10.1007/978-981-33-4597-3_37; https://doi.org/10.1007/978-981-33-4597-3_37
• Institution: Universiti Malaysia Pahang Al-Sultan Abdullah

In multilevel sentiment classification task, there is a challenging task of limited coherence, contextual and semantic information. This paper proposes a new hybrid deep learning architecture for multilevel text sentiment classification with less training and simple network structure for better performance and can handle the implicit semantic information and contextual meaning of text. In this research the proposed hybrid deep neural network architecture made with Bidirectional Gated Recurrent Unit (BiGRU) and Bi-Directional Long Term Short Memory(BiLSTM) of Recurrent Neural Network (RNN) for multilevel text sentiment classification and this performs better with higher accuracy than other methods compared. This proposed method BiGRUBiLSTM model outperformed the traditional machine learning methods and the compared deep learning models with about average of 1% margin accuracy on different datasets.