Research on Sentiment Analysis of Two-way Long and Short Memory Network Based on Multi-Channel Data

In recent years, with the rapid development of deep learning and the Internet, Weibo and e-commerce platforms play an important role in people's daily lives. User data and Weibo comments on e-commerce platforms hide rich emotional information. At the same time, these massive amounts of data also attract more researchers to participate in the research of short text data such as online shopping and Weibo comments, researchers can analyze these data to help businesses increase sales and improve product quality. On the other hand, it also provides a reference for consumers to fully understand the attributes of commodities. How to dig deeper into the potential emotional information of short texts is the most important thing. The emergence of neural networks has pushed natural language processing sentiment analysis to a hot topic, this article uses a combination of convolutional neural network and attention mechanism to solve this problem. This paper proposes a two-way long and short memory network model based on multichannel data. The model first uses shallow learning features such as emotional part of speech extracted from short texts to map to high-dimensional space to form 3 high-dimensional feature vectors, combine the three feature vectors with word embedding and input them into BiLSTM form three channels, the attention mechanism is added to the three fully connected layers, and finally input to the SoftMax classification layer. The experimental results show that compared with the traditional text convolutional neural network, the method proposed in this paper has a significant improvement in the accuracy of short text sentiment analysis tasks.