On Subjective Test's Automatic Scoring System In The Field Of Rail Traffic Signal Based On Ontology And Syntactic Structure Analysis

Along with the computer technology,and internet communication technology etc.,distance education has become the most promising new teaching pattern.As an important part in the distance education system,automatic scoring plays an important role in ensuring the fairness of scoring and reducing the workload of examiners,so the development of automated scoring technology has speeded up the pace of distance education towards intelligent education.In the rail transit system,the distance education system is applied to the induction training and continuing education of the personnel in every department.Moreover,rail transit signal is a compulsory knowledge of rail transit electrical staffs,so it is often used as the assessment content of the electronic service system training in all stages.At present,the automatic scoring system of objective test has been perfected,and almost all large-scale examinations' the objective test of reading task in China are completed by automatic scoring system.In comparison,the subjective test's automatic scoring system is in the research stage.Although the domestic related research has made progress,the system is still not mature enough,because there is a large error in semantic analysis,the understanding of the professional concept in a specific field is not accurate,the system similarity calculation does not consider the semantic relationship of the text etc.,and no scholars have carried on the research of the automatic scoring system in the field of rail transit signal.Based on further study of the development trend of the automatic scoring system and a detailed analysis of the advantages and existing problems of the existing system,this paper proposes an automatie scoring system of subjective test in rail transit signal based on analysis of ontology and syntactic structure.The major work of this paper includes the following aspectse:First,the rail transit signal ontology is constructed.In this paper,the rail transit signal ontology is constructed and introduced into the system,which provides the semantic knowledge source in rail transit signal.Second,a text preprocessing method is proposed.Using natural language processing technology,this paper puts forward the method of relation word to preprocessing,in order to preprocess the answer text,which includes:text clause participle,lexical annotation,syntactic structure analysis,anaphora resolution,processing of dependency among word pairs,etc.After preprocessing,the text is transformed into a set of relational word pairs that can reflect the semantic relationship of each component in the text.Third,a new text similarity algorithm and scoring algorithm are designed.Aiming at the characteristics of system data and existing systems,combined with the characteristics of professional words in the field of rail transit signals,this paper designs the similarity algorithm and scoring algorithm for words and relational words suitable for reading tasks in the field of rail transit signals.The similarity algorithm and scoring algorithm take into account the similarity of words in the field of rail transit signal,the similarity of common words,the similarity of syntactic structure and the similarity of sentence patterns,and solve the problems of low accuracy of the existing reading system in the field of rail transit signal and the syntactic relationship of the system algorithm without considering the text.Fourth,the software system is developed and several aspects of the system are tested experimentally.According to the designed algorithm,the automatic reading model is constructed,and the software system is developed based on the project of"ATS Teaching Simulation System of Changzhou Metro Line 1 ",and has been tested and validated.There are 23 figures,19 tables and 52 references.