Text classification
Text classification is an important task in natural language processing, which is often applied to sentiment analysis, news filtering, spam detection, and other scenarios8. Text classification uses features to represent raw text and provides them as inputs to downstream classifiers. The most commonly used representation is Word2vec9, which uses low-dimensional dense word vectors to represent words, but it ignores the semantic relationship between words, so it faces problems such as data sparsity and polysemy. In recent years, deep neural networks such as convolutional neural network (CNN) and recurrent neural network (RNN)10have been applied to extract contextual information and semantic representation from text. The results show that the performance is better than the traditional method. Kim et al . achieved good results in sentence classification by using different filters to extract multi-granularity feature sentences11. Sinha utilized bidirectional long and short-term networks to convert words into context embedded representations12, enabling the network to learn contextual information in statements.