Text classification method based on LSTM

Abstract

With the rapid development of Internet technology and the explosive growth of social media, a large amount of information continues to be generated, of which the amount of text information is the largest. The main feature of various Chinese short text messages such as news headlines and instant messages is sparsity, which is only composed of a few to dozens of words, and the content of effective information packets is very small. As a result, the samples with sparse features and high feature set dimensions are difficult to provide key and accurate features for text classification learning. This paper mainly studies the application of deep learning in the field of Chinese text classification, and proposes a text classification model based on word level and character level mixed features.