Research on New Word Recognition Based on Weibo Information and SVM

Abstract

In order to effectively recognize Weibo new words, this paper proposes a Weibo new word recognition method based on Weibo information and SVM. Firstly, positive and negative samples are extracted from the Weibo corpus and the training corpus annotated with part of speech. Next, the word features of these samples are vectorized and trained through support vector machines to obtain support vectors for new word classification. Finally, it is vectorized and inputted into the trained support vector machine classifier to obtain the recognized Weibo new words. By comparing the experimental results, the optimal fusion feature combination and SVM kernel function are obtained.