of the teacher, this assessment can be automatically transferred to one of the systems
of knowledge assessment (Bologna, five-point).
If the teacher put some initial complexity of all the tasks of the block (
0
,=
1,
̅̅̅̅̅̅) then the system must consider, some heuristics when choosing the next question:
• Complexity of the question the system chooses according to the average level of
student's knowledge (
∈[0,1]);
• If the student gives the correct answer, then the complexity of the question
decreases, and vice versa;
• The dynamics of the values of the complexity of tasks depends on the number
of students passing the assignment and on the average level of knowledge of each;
• If a student with a relatively high level of training gives the correct answer, the
complexity of the question decreases less than for a student with a low level of
knowledge;
• The minimum complexity value should be greater than zero
• The maximum complexity value is not more than one;
• With a large number of students, the complexity of the question will come to
some constant value;
After the testing (when the last student completed the tasks) the system lists the
cost of the questions and writes the new values to the database. If in the assessment of
knowledge participated n students and there were m questions, then the correction of
complexity will be calculated according to the formula:
=
−1
+
(
0
,,
−1
,), =1,
̅̅̅̅̅, ,=1,
̅̅̅̅̅̅
Where
- the complexity of the i task when passing j student;
- function
according to which the correction of the estimation takes place;
−1
total score j-1
student; Z - a certain value, which depends on the rejection of this answer from the
correct one.
Depending on the type of question, this formula undergoes some changes, but its
essence will not change. In general, there are 8 types of questions.
CONCLUSION
The article deals with the types of systems of electronic education and the
application of methods of adaptation adjustment of complexity of tasks during the
evaluation of students using test-evaluation electronic systems (on the example of the
SMPR system and the corresponding website). System developers plan to continue
improving these algorithms for the system and spread the experiment to other
disciplines, in particular, numerical methods and research operations involving a wider
range of students for testing. As a laboratory work, the students of the Faculty of
Computer Science and Cybernetics developed and improved a program product for
- 1848 -