Authors
Polyakov V.
|
Degree
|
Associate Professor of Perm State National Research University |
|---|---|
|
E-mail
|
polyakov@mpm.ru |
|
Articles
|
Fuzzy inference alghoritm for multi-factor student assessment ratingPresently estimating students rating envolves a lot of different factors which have to be taken into
account. The rating mark has to follow some regulating documents which enclose recommendations
that are formulated as complex of inference rules that are formulated in wordy («linguistic») form like
«if… then». It is evidently that in these conditions experts are needed for student estimation. As a result
the mark is far from having a single meaning. To overcome this difficulty we propose to use the
methods of fuzzy logic. It was found that for implementation of fuzzy inference algorithms the matrix
representation of fuzzy variables [1] is most effective. In this representation the usage of logic
rules like modus ponens is equivalent to solving of some system of linear equations with the usual
existing conditions of their solution. The comparison of matrix and traditional approaches on different
software and hardware platforms (including embedded and mobile systems) was also realized.
Main advantages of matrix method for are low memory requirements, shorter and easier development.
As a result the proposed algorithm enables automate the procedure of student rating estimation
which sufficiently taking into account expert opinion. In other hand, it makes the student mark
transparent in difficult cases.
Read more...
|
