Model development for results prediction of student testing in computer science based on the results of monitoring

Testing is the most widespread and popular way for quick estimation of student knowledge. The application of testing in the educational process at Bryansk state university at the Department ‘Computer science and Software support’ in the course ‘Computer science’ made it possible to decrease significantly labor coefficient of knowledge monitoring, heighten the interest in the discipline, as well as increase the level of training. In the course of testing application and statistical data accumulation it was occurred an idea about evaluation of students’ level of training in computer science. Preliminary survey of the training process for final testing showed the dependence between quantity of successful attempts in passing tests and their results. In this connection we have developed functional simulators for testing process and defined the prediction phase. The purpose of the prediction phase is to inform the student about his/her training level and expectancy for a successful result. That allows the student to set a goal on his/her own and strive to achieve it. On the other hand, the teacher has an opportunity to monitor the student’s level of knowledge and assist him/her in proper time. For correct prediction of training level coefficient, different approaches for statistic evaluation of the data obtained during the process of testing were studied. The approach analysis showed the possibility of using the method based on the theorem of joint event. Since the academic course is arranged from simple to complex subject-matter, the low coefficient can influence on the further material. For carrying out the preliminary computation, the software for dealing with electronic worksheets was used. The essential estimations were carried out as well as various combinations in initial data were considered. The initial data were obtained from electronic educational complex “UMK-A”, which was developed by the authors of this paper. The complex has been used in the educational process for the last several years. The results of the preliminary estimations proved the application of this model. The findings showed the conformity to the expectations obtained by empirical method. The decision was made to integrate the algorithm developed for prediction of training level into the electronic educational complex “UMK-A”.

Keywords

Authors

References