Zhizhong Liu*, Bin Zhou, Lingqiang Meng, Guangyu Huang
Intelligent Automation & Soft Computing, Vol.37, No.2, pp. 1963-1981, 2023, DOI:10.32604/iasc.2023.038835
Abstract Recently, multimodal sentiment analysis has increasingly attracted attention with the popularity of complementary data streams, which has great potential to surpass unimodal sentiment analysis. One challenge of multimodal sentiment analysis is how to design an efficient multimodal feature fusion strategy. Unfortunately, existing work always considers feature-level fusion or decision-level fusion, and few research works focus on hybrid fusion strategies that contain feature-level fusion and decision-level fusion. To improve the performance of multimodal sentiment analysis, we present a novel multimodal sentiment analysis model using BiGRU and attention-based hybrid fusion strategy (BAHFS). Firstly, we apply BiGRU to learn the unimodal features of… More >
Login
Register
