MODELLING AND OPTIMIZATION OF THE STUDENT’S RATINGS

Recently higher educational institutions are developing the rating system. Due to the rating can be automated grading, making this assessment more objective, independent from the teacher. Rating system allows to plan and predict for student the quality of their education. Rating system is the algorithm, which gives for each student at the end of training a certain score, expressed by a number. By means this ball can be exhibited final evaluation, usually on a five-point scale. To build a rating system, we must 1) to specify the number of indices expressing the student’s ratio to learning (usually grades for the current tests, calculation-graphic work for extracurricular execution, theoretical colloquium, indicators of activity of the student in learning: working by board, consultation with teachers, attendance), which formed the rating of the student; 2) to put into correspondence with each indicator three numbers: minimum, normal and maximum levels; 3) to allocate among the indicators the most important, such that if the student does not reach the normal level at the specified index, it does not receive a satisfactory final evaluation; 4) to construct an algorithm, by means which student’s rating is formed. on the basis of specified levels. The last task is the most difficult of the above and many researchers are solving it. It is known that one of the tools for exploring the many segments of the educational process is statistical modeling. Thanks modeling we can - to optimize, improve the quality indicators of educational systems, - to explore the impact of different factors on the qualitative aspects of the educational process. In this paper the author offers methods of statistical simulation some of the indicators forming the rating: current assessment of control works, assessment of calculation-graphic work, the evaluation of the theoretical colloquium, the activity of the student in the study of the course , attendance. In addition, we offered the statistical method of selecting a function of the rating, which allows to determine the most effective function. This method is based on the use of mathematical model of Rasch. To solve the problems we have developed a computer program that simulates the process of ranking by means to the Sablin’s algorithm. Through this program we have showed the efficacy of Sablin’s rating function in relation to the selected control piecewise linear function.

Keywords

Authors

References