Judgments on the Pros
The content provided in the paper adequately evaluated and compared sophisticated machine learning applications that can be utilized in solving the problem of predicting the exam performances of students. This involves identifying students highly likely to drop out of the course and predict their achievements in the future (Tomasevic, Gvozdenovic & Vranes, 2020). In the classification and regression functionalities, the highest level of accuracy is obtained by artificial neural applications. This was achieved through use of student engagement information and the results of their past performances.
When determining the potential of the performance prediction application of the student exam, sufficient acquisition functionalities, and the interactions of the students are valuable. This will ascertain that adequate information can be accessed for assessment. The research showed the importance of having an Open Course Ware platform since it can be used in hosting educational information. Within an application, the informative content is designed and highlighted courses and will represent a daunting endeavor in the implementation of learning analytics (Tomasevic, Gvozdenovic & Vranes, 2020).
However, the researchers indicated that when such critical information is analyzed with appropriate data mining techniques, an excessive amount of data can be determined as a valuable resource for educators. Effectively utilizing education and digital technology on a similar platform will influence the learning and teaching cultures that allow data access, primarily from online learning environments. As a result, these can be essential in improving the learning experiences of students.
Cons and recommendations
The researchers did not include emerging educational data mining techniques such as matrix factorization. If the study used such a method, it could have focused on the deployment of content recommendation systems. The will succeed if an adequate dataset of learning data as such information contains various teaching courses across a large number of students. Once this is made possible, it leads to an assembly of a student-course matrix that is reliable enough for the evaluation by matrix factorization and also other suitable techniques.
The study did not fully determine the level of success of the performance prediction of the student exam. It did not analyze the manner that its applicability can be limited to particular learning environments. Through utilizing such exam performance prediction with the application of adequate data acquisition functionalities whereby student interaction is necessary, this objective will be achieved due to a sufficient amount of data that can be analyzed.
Discussion
How can the issue of data sparsity be solved, considering it directly affects the effectiveness and accuracy of learning analytics?
Data sparsity occurs since there is inadequate information concerning the activities and engagement of students. This is because numerous students do not actively engage with learning mediums, such as offline learning materials. This leads to only a small number of students interacting with the learning platform and provides essential data than can be assessed further. This challenge can be solved through the application of the necessary mediums that will trigger the learning activities and enhance the process of acquiring valuable data. This can be achieved by emphasizing to the students that interaction with the learning platform is mandatory, which ultimately will increase the frequency of interactions. This induces engagement by these students, and therefore, relevant data can be acquired from a large population of students.