Tuesday, February 19, 2008

PCSC 2008: Dr. Elmer Maravillas

8th Philippine Computing Science Congress (PCSC 2008)
23 - 24 February 2008
University of the Philippines-Diliman
Quezon City, Philippines
Organized by: Computing Society of the Philippines (CSP)
Tel. +63 2 7090907, +63 2 4266125
E-mail: computingsoc@gmail.com
Website: http://www.csp.org.ph/
=================================================


Expert System’s Evaluation of Student Learning Performance

Elmer A. Maravillas, PhD
Head, Research and Development Coordination Office (RDCO)
Cebu Institute of Technology (CIT), 6000 Cebu City, Philippines
Email: elmer2717@yahoo.com


ABSTRACT

This paper presents an evaluation strategy for student’s learning performance. It uses the expert system’s approach to reasoning with uncertainty. Given the answer sheets of students and the point assignment to each question of an examination, the strategy computes for the total scores.

These scores will be ranked from the highest to the lowest. Students clustered with the same scores will be re-ranked by considering the difficulty, complexity, and significance of each question in the examination.

The individual rank of students that belong to a cluster will be determined by utilizing adjustment values. Each question will have a distinct adjustment value which is inferred using a forward-chaining inference procedure with fuzzy logic as the backbone for reasoning under uncertainty considering the degree of difficulty, degree of complexity, and degree of significance.

The sum of difference of students belonging to a cluster will be determined using the adjustment values, points assigned to each question, and the scores obtained by the student for each question. By sorting the sum of difference of each student in descending order, the true ranking of the students will be revealed. Experimental results show that this method is able to really evaluate the true and distinct learning performance of students.

No comments: