Resumen de Analysing the variation of metadiscourse verb patterns in English academic papers from within and between disciplines

Yibo Li

• English is a universal language and is widely used. In order to deepen the cognition of language, this paper proposesto analyse the variation of metadiscourse verb patterns in English academic papers from within and between disciplines.By comparing the metadiscourse chunks used in various disciplines, it is concluded that there are obvious differencesin the verb types of the subject papers. Analysing various identification algorithms of verb types, it is obtained that thecombination of the word movement distance (WMD) model and the R&L density peak clustering algorithm is the best.Using the R&L density peak clustering, identification parameters are easy to determine, and when combined with theWMD model to calculate the similarity of words, it improve the accuracy of verb pattern clustering. By comparison, it isproved that the accuracy of the algorithm combining the WMD model and R&L density peak clustering reaches 30.06%,and the effect of identifying verb pattern variation is the best