چكيده لاتين
Today, online education is more welcomed by students and professors than ever before. Compared to traditional education, online education provides more accessibility and higher quality. Personalization is another trump card for online education. By personalizing education, it is possible to take advantage of the difference between students in a way so that in addition to increasing their learning rate, the student dropout rate as a measure for measuring the quality of educational systems is also reduced. One of the factors affecting the drop rate of students is their motivation. Therefore, the aim of the research is to personalize the order of presenting the content in such a way that the studentʹs motivation is also maintained. It can be expected that the personalization of education, while maintaining student motivation, will lead to a decrease in their dropout rate.
By considering studentsʹ motivation in personalizing the order of providing online education content for them, we can expect more correct suggestions to be provided. Paying attention to this psychological feature is an idea that is considered in this research. Different researches have shown that determination and motivation of students have close definitions and have a significant effect on each other; Therefore, by measuring studentsʹ determination, we can get closer to the definition of their motivation, and then, specifically by addressing motivation, we can obtain a more accurate and reliable method for personalizing online education.
In order to include motivation in personalizing the order of teaching content, the proposed method is developed in two steps. The first step is to determine whether students are determined or not. For this purpose, a clustering model from the subset of machine learning has been trained, which obtains the studentʹs determination based on the activity history of the student. In the second step, based on existing researches in the field of psychology of education and motivation, methods for quantifying motivation were implemented; Then, using reinforcement learning, an intelligent agent was trained to suggest personalized educational content to each student by observing and modeling the educational paths of different students. The evaluations and obtained results show that by using machine learning algorithms, determined and non-determined students can be clustered with a silhouette score equal to 0.5020. The results prove that despite the qualitative and abstract nature of the concept of motivation, it can still be quantified using different techniques and by using it in reinforcement learning, with higher accuracy and precision, the unique educational path can be suggested to each student. The average click rate of the proposed method of this research for personalizing education is 0.24.
