News Text Classification Based on Improved Bi-LSTM-CNN

Abstract

The traditional text classification methods are based on machine learning. It requires a large amount of artificially labeled training data as well as human participation. However, it is common that ignoring the contextual information and the word order information in such a way, and often exist some problems such as data sparseness and latitudinal explosion. With the development of deep learning, many researchers have also been using deep learning in text classification. This paper investigates the application issue of NLP in text classification by using the Bi-LSTM-CNN method. For the purpose of improving the accuracy of text classification, a kind of comprehensive expression is employed to accurately express semantics. The experiment shows that the model in this paper has great advantages in the classification of news texts.