Examine study pattern on selective cross join data using bootstrap

A learning analytics model uses students' academic records to recommend study paths based on the their academic performance. It also encourages students to improve their performance on the subjects in which they had a lower grade. Subsequently, the process of implementing a learning analytic system for study path recommendation can be carried out by developing a knowledge base model using selected cross-join data. In this study, the selective cross-join technique, which was implemented using the bootstrap validation method, was examined. Furthermore, the data used are drawn from student records from the previous two academic years that have already undergone pre-processing to eliminate any newly added courses, since there would not be much to learn from them. The validation process, which took 10 iterations, was carried out using the bootstrap method and the result for each iteration was evaluated using 1 - Root Mean Square Error. The lowest, highest, and average accuracies obtained from all 10 iterations were 69.2%, 92.3%, and 84.69%, respectively. This inconsistency indicated that the process may have been misinterpreted without taking into account any noise that might have been replicated in the data.