Özet:
Prediction of student performance is one of the most important subjects of
educational data mining. Artificial neural networks are seen to be an effective
tool in predicting student performance in e-learning environments. In the studies
carried out with artificial neural networks, performance predictions based on
student scores are generally made, but students’ use of learning management
system is not focused. In this study, performances of 3518 university students,
who studying and actively participating in a learning management system, were
tried to be predicted by artificial neural networks in terms of gender, content
score, time spent on the content, number of entries to content, homework score,
number of attendance to live sessions, total time spent in live sessions, number
of attendance to archived courses and total time spent in archived courses
variables. Since it is difficult to interpret how much input variables in artificial
neural networks contribute to predicting output variables, these networks are
called black boxes. Also, in this study the amount of contribution of input
variables on the prediction of output variable was also examined. The artificial
neural network created as a result of the study makes a prediction with an
accuracy of 80.47%. Finally, it was found that the variables of number of
attendance to the live classes, the number of attendance to archived courses
and the time spent in the content contributed most to the prediction of the
output variable