English Linguistic Term Extraction and Classification Strategies under The Influence of Network Language

The network language used in online chat has its commonness and characteristics. However, the level of word segmentation in the network text is not satisfactory, which puts forward a new topic for the current research on term extraction. This paper summarizes and analyzes the characteristics of the network text. Based on the conditional random field model and the combination of word tagging, the tagging set is improved, and the feature template suitable for the network text is adopted to improve the classification ability of the network text. We use the method of combining statistics and rules to recognize the Internet text named entity, normalize the text, and use CRF model to train and predict the text. Integrated with the matching rules, we modify the recognition results to achieve better results. The experimental results show that compared with the traditional text regularization and SVM classification methods, the performance of the proposed scheme is significantly improved.