Using Random Forest on Formative Assessments to Enhance the Quality of Teaching and Learning in Engineering Education

Abstract

With the popularization of engineering education professional certification, the goal of engineering education is more and more oriented to students' comprehensive ability, which also makes formative assessments more important in the process of talent training. As for a course, not only the assessment results but also the learning process of students should be paid more attention. By setting up rich and colorful process learning tasks, students can better achieve the course objectives. However, for the development of process learning, two problems require more attention. One is how to evaluate students' enthusiasm and learning effect and give an effective early warning, and the other is to analyze the effect of different learning tasks on students' final achievement of learning goal. In this study, on the basis of process learning historical data, a student process data evaluation and analysis model based on random forest is established. The data coming from Superstar Learn Platform is the process learning data of more than 300 students in a University. The analysis results show that appropriate curriculum design, course assignments and chapter tests have a significant impact on promoting students to achieve the course objectives. Besides, the timeliness of students' completion also plays a significant role. This analysis model can help students correct their learning attitude, help teachers better understand students and adjust teaching plans.