Research And Implementation On Part-Of-Speech Tagging In Automatic English Essay Scoring

With the deepening of the opening up, the number of English learners increased dramatically, English learners articles related analysis seems increasingly important.Automatic English essay scoring system utilization of natural language processing, statistical, corpus linguistics and information retrieval techniques. Aims to provide an excellent way to learn for Chinese English learners, improve writing skills of learners. Automatic English essay scoring system is designed and implemented for Chinese English learners.Part-of-Speech tagging for Chinese English learner language is the base of automated essay scoring system. Although researchers have done a lot of fruitful Part-of-Speech tagging researches, most of them are focused on the English essays written by native speaker, and there is no research about essays of Chinese English learner. We annotate essays of Chinese English learner manually, and propose an algorithm of Part-of-Speech tagging for Chinese English learner language. This method uses a lot of rich features, such as unsupervised word cluster, unsupervised tag dictionary and phonetic normalization. With these features, our system outperform state-of-art tagging results on the manually annotated corpus, tagging accuracy is improved from94.49%to97.07%. By improving speech tagging system, the syntax error detection of automatic English essay scoring system has been improved.