The Research And Implementation Of Automated Assessment System In C Language Program

Automatic assessment of student programs can greatly reduce teacher’s work andimprove the consistency and accuracy. However, it is also a very challenging topic since itinvolves automatic test data generation, proving program correctness and so on. There isalready exist a number of different automatic scoring systems at home and abroad, the mainscoring methods are divided into dynamic and static assessment. Different systems fordifferent user groups, and they use different scoring methods and metrics, so we will get adifferent evaluation results. This issue is designed for automatic scoring non-computerprofessional students’ C language programs, the objectivity of the automatically scoringresults is the main problem what we wants to solved, the subject has theoretical andpractical significance.In this paper, it put forward a solution to automatic scoring student programs based ondynamic and static assessments, it is combined the dynamic and static methods together.This process is divided into three stages: Firstly, check the result file of the program, if itnot exist, then compare the source code directly, if there is a result file, then run the file first,and compare the computational results; Secondly, we find the source code for simplelexical analysis, extract the keywords of the student and template programs, given scoresaccording to the string matching algorithm; Finally, we use the standardized rules of controlstructure to standard the student and template programs, on this basis, compare thestandardized program from student and template, then given the score according tocalculate the similarity of the standardized programs. At last, the final composite scores aregiven according to the weights of the three scores.Finally, we adapt the solution in automated assessment C language programs. Theexperimental results showed that: what keyword we set for the programs, how we set theexam questions and the method that combination of dynamic and static assessment in theprogram, are all affected the automated assessments. In this paper, we use the experimentdata verified the objectivity of automatic scoring.