A Brief Survey of Text Classification Methods

Text classification is a long-standing research spot in natural language processing (NLP), which greatly facilitates various NLP tasks. With the rapid development of Internet and social media, the need for effective data management also makes text classification become more and more important. Its research methodology has also developed from the early shallow classification based on handcrafted features to the deep language modeling with various neural networks. In this paper, we first give a brief introduction to its background and definition, and then describe its representative methods at different stages of development. Meanwhile, we also introduce the recently popular pre-trained language models on large-scale corpus, and discuss their applications to text classification. Overall, this paper gives a quick guide to understanding text classification research.