A Corpus-based Study Of Business English Writing Automatic Assessing System

As its high validity, the writing test is adopted by almost any kind and any scale language test. But, the reliability of the writing test rating is broadly questioned. As the development of the computer technology and the integration of the language test and the statistics, the computer-assisted writing assessing systems appeared and adopted by many authentic large-scale language test home and abroad. The already existing writing automatic assessing systems are almost based on the training sets which are trained by the human raters. So, essentially, these automatic assessing systems are the imitations of the human rating. In this thesis, take the business English (one of the most important branch of the ESP study) as the breakthrough point, base on the authentic data derived from the relevant corpora, the author is going to prove the feasibility of the corpus-based business English writing automatic assessing system.