Evaluating Distance Learning Students’ Performance by Machine Learning
It is essential to predict distance education students’ year-end academic performance early during the semester and to take relevant measures on the basis of such prediction. In this way, it could be possible to enhance students’ academic performance and thus the quality of education they are provided with. The present study is forced on designing a new mathematical model to predict distance education students’ year-end academic performance in reference to the 6-week data kept in the learning management system. Particular fuzzy models were formed on the basis of classical fuzzy and learned opinion, and the membership function ranges for the fuzzy model were optimized through a genetic algorithm. Next, a hybrid model was formed with fuzzy logic using clustering methods. The data were collected on Moodle, an open-source learning management system. The model was based on the data on a total of 218 students registering for the course Basic Computer Sciences during the 2011-2012 Academic Year. In addition, one more dataset, which was concerning another 95 students registered for the course during the 2012-2013 Academic Year, was used as verification data. The input data were comprised of five components, namely the last time when students logged on to the system, the frequency at which they logged on to the system, the amount of time spent online in the last session, one quiz administered in Week 4, and one midterm exam administered in Week 8. The models formed on the basis of these data were used to predict students’ academic performance and to make particular comparisons.
ŞİŞMAN Burak
Danışman : Prof. Dr. Sıddık B. YARMAN
Anabilim Dalı : Enformatik
Programı : Enformatik
Mezuniyet Yılı : 2014
Tez Savunma Jürisi : Prof. Dr. Sıddık B. YARMAN
Prof. Dr. H. Ali ÇIRPAN
Prof. Dr. İlhan KOCAARSLAN
Prof. Dr. Sevinç GÜLSEÇEN
Doç.Dr. Ümit GÜZ
Dostları ilə paylaş: |
