Article Preview
TopIntroduction
In recent years, due to the increasingly heavy task of English learning for middle school students, there has been an urgent need to rapidly improve students’ English writing ability (Kong & Lee, 2021). When scoring English compositions, due to significant differences among students, teachers need to spend a lot of time reviewing and explaining and cannot balance the discrepancies in student abilities. Teaching students according to their aptitude results in students being unable to receive timely feedback and engage in large-scale writing exercises (Albiansyah & Minkhatunnakhriyah, 2021). At present, the development of automatic scoring (AC) systems for objective questions in foreign countries is relatively mature, but the research on scoring systems for subjective questions such as compositions is not yet as developed (Zhao, 2021). The Link Grammar (LG) analyzer is a key component of an AC system. It can quickly evaluate English sentences and is robust. However, it only targets complete sentences, and the number of dictionaries is limited, making it difficult to accurately recognize abbreviated and complex sentences. It is easy to misjudge correct sentences, which affects the accuracy of the entire composition score (Chen & Liang, 2022). The N-grammar model can not only filter out misjudged sentences using binary standards but also improve the output results. To design a more accurate composition scoring model, this research introduces an N-grammar model, combines LG with the N-grammar model, and analyzes and evaluates grammar in composition; it then uses information entropy (IE) to judge composition level, synthesizes the grammar weight of each sentence to score grammar, and finally calculates composition score by combining composition level with overall grammar weight. The model aims to shorten the time for teachers to evaluate compositions, improve their work efficiency, and increase opportunities for students to practice English writing. Detailed analysis of grammar in research can also help students to improve their grammar and improve their English writing ability. The innovation of this research lies in the design of a human-machine integrated English composition intelligent scoring (CIS) system, which improves the accuracy of composition scoring. The research content is divided into four parts: The first part is a brief introduction to the relevant research on cloud classrooms and scoring systems; the second part first introduces two types of grammar detectors in detail and then constructs a CIS model based on these grammar detectors; the third part is the application of the CIS model, during which performance testing and comparative analysis experiments are conducted, and the results of this application; and the fourth part is a summary and outlook of the research content.