Construction of Intelligent Evaluation Platform Based on Random Matrix iWrite College English Writing

Abstract

The advantages of iWrite English writing teaching and marking system 3.0 and the case study of English writing in higher education based on this system have a positive effect on forming a new model of English writing teaching in higher education under the background of Internet + and improving and solving the main problems of English writing teaching in higher education. Although the current intelligent marking systems have made a lot of achievements, they have not fundamentally solved the problem of the rationality of intelligent marking of subjective questions. In order to better perceive the sense of speech in English essays in depth and improve the rationality problem of intelligent marking, this study proposes a quantification of N-element sense of speech value based on correlation analysis and a scoring fitting algorithm for English essays based on rationality enhancement. The quantification of perceptual value is done by obtaining the N components of the composition and calculating their support in the corpus. In the daily teaching work, teachers should actively cooperate with the marking system to improve their own teaching level and explore ways and methods of students’ writing training, so as to cultivate more English talents. In addition, word features, sentence features, and chapter structure features were extracted from the essays to fit the English essay scores. Since not all students were able to complete the essays according to the requirements of the questions, a streaming scoring model was used to separate the normal essays from the low-scoring essays. The low-scoring essays were scored using the k-nearest neighbor algorithm, while the normal essays were fitted to the test takers’ scores using the support vector regression algorithm. Statistically, it was found that the essay scores also showed a certain normal distribution. The standard support vector regression algorithm is prone to data skewing problems, so this study addresses this problem by using a rationality enhancement method, which gives a corresponding penalty factor according to the distribution of the data set.