AIoT based sentiment analysis via multi-scale context perception to enhance linguistic teaching

With the rapid development of the Internet of Things (IoT) and artificial intelligence technology, real-time text sentiment analysis plays an important role in online education. Due to the limited resources of clients, existing deep networks cannot be directly deployed in the edge node. In addition, deep convolutional networks cannot fully utilize contextual information. In order to resolve these issues, this paper proposes a multi-scale context-aware text sentiment analysis system based on cloud computing, in which a bidirectional long short-term memory network (BiLSTM) model is deployed in the cloud server. The BiLSTM model can fully explore the contextual feature information of the text stream in online education. The real-time text data are collected through terminal nodes, such as pad or computer, to send the cloud server. The experiments utilize three public text datasets to simulate the input of terminal nodes. The results show that the proposed system shows better accuracy than previous models and can return the emotional status in time.