Research On Network News Text Classification Model Based On ERNIE-CBi-Att

With the development of the Internet and the advent of the Big Data era in the 21 st century,the number of Internet users in China has exceeded one billion.Nevertheless,alongside the increase in internet users,there is a faster growth rate of network information data with a larger increase,among which network news is is a significant means for internet users to acquire external information and gain knowledge,as well as enhance their understanding of the world,which is favored by many netizens.Network news primarily exists in textual form.Nowadays,the news style has moved from a fixed form to integration and diversity,and presents a hybridness feature,which makes it increasingly challenging for users to promptly access the news information they desire to acquire.The task of text classification in natural language processing is to solve such problems.Text classification technology has developed from the early expert manual processing method to traditional machine learning and now to deep learning,and has formed a relatively complete classification system.Based on this,many good performance network news text classification models appear in people’s sight.The emergence of Web news text classification technology enables users to better classify news information to read,so as to achieve more effective resource saving.In addition,this technology is of great importance in applications such as providing users with personalized news recommendations,categorized navigation,monitoring public opinion and spam filtering.At present,as a technology developed from traditional machine learning,deep learning can better describe the characteristics of text data and process text data efficiently,consequently,the academic community has devoted considerable attention and conducted extensive research on it.In this paper,based on the deep learning model,a network news text classification model based on ERNI-CBI-ATT is constructed.Specifically,the model performs text representation operation through ERNIE pre-trained language model.Then,the convolutional neural network and the bidirectional gated recurrent unit network are concatenated as the text feature extraction layer.Then,an attention mechanism is added to weight the feature vectors.Finally,the final feature vectors are fed into the classifier to achieve the classification of online news texts.The results show that the ERNIE-CBi-Att model constructed in this paper has a slight improvement compared with other models in four evaluation indicators in commonly used text classification.