| Writing,as an important and necessary content in the examination of students’ English ability,can comprehensively reflects the students’ mastery and application ability of English vocabulary,sentence patterns,grammar and other indicators.Therefore,in the English test,writing scores account for more weight than other question types.Compared with other types of questions,the task of scoring essays is more arduous and time-consuming,and because the evaluation criteria of essays are relatively flexible,the objectivity and fairness of the scoring are more difficult to guarantee.At present,researchers have developed many automatic scoring systems for English writing,but few can be used in actual examinations.This paper studies the related results of the automatic scoring technology of English writing.The text characteristics of the essay can be divided into two levels: language characteristics and content characteristics.The language characteristics represent the language information in terms of vocabulary and syntax,and the content characteristics represent the content of the essay.Through modern natural language processing technology,the text is segmented,sentenced and syntactically analyzed to obtain the shallow features of the composition.The local semantic information of the composition is extracted through the convolutional neural network and the global semantic information of the composition is extracted by the long and short-term neural network.To obtain the deep features of the composition,and then perform feature fusion to build an automatic scoring model for English essay.After a large number of experimental demonstrations,the results show that the model can effectively completes the task of automatic scoring of English essay,which proves the feasibility of the method.This paper refers to the actual situation of the English writing scoring in China’s senior high school entrance examination,carries out demand analysis,system design and system implementation,and builds an automatic English essay scoring system that can be used to assist in the scoring of English examinations.Through a series of system tests,the system has shown more reliable practicability in terms of function and performance,and meets the requirements of the automatic scoring of English essays in Senior High School Entrance Exam. |