From Transcripts to Insights for Recommending the Curriculum to University Students
- PDF / 2,920,257 Bytes
- 14 Pages / 595.276 x 790.866 pts Page_size
- 15 Downloads / 176 Views
ORIGINAL RESEARCH
From Transcripts to Insights for Recommending the Curriculum to University Students Thong Le Mai1,2 · Minh Thanh Chung3 · Van Thanh Le1,2 · Nam Thoai1,2 Received: 29 May 2020 / Accepted: 14 September 2020 © The Author(s) 2020
Abstract Student data play an important role in evaluating the effectiveness of educational programs in the universities. All data are aggregated to calculate the education criteria by year, region, or organization. Remarkably, recent studies showed the data impacts when making exploration to predict student performance objectives. Many methods in terms of data mining were proposed to be suitable to extract useful information in regards to data characteristics. However, the reconciliation between applied methods and data characteristics still exists some challenges. Our paper will demonstrate the analysis of this relationship for a specific dataset in practice. The paper describes a distributed framework based on Spark for extracting information from raw data. Then, we integrate machine learning techniques to train the prediction model. The experiments results are analyzed through different scenarios to show the harmony between the influencing factors and applied techniques. Keywords Educational data mining · Prediction · Student performance · Machine learning · Distributed system · Spark
Introduction Education data mining (EDM) is a research field which concerns data-mining techniques to analyze patterns from data in educational context [33]. Online learning systems such as Learning Management Systems (LMS) [6], Massive Open Online Courses (MOOCs) [25] have become more and more popular in higher education institutions with the advance of current technology. Sometimes, it is a requirement every student needs to participate as a course rule, an external factor such as a global pandemic where all universities are forced This article is part of the topical collection “SoftwareTechnology and Its Enabling Computing Platforms” guest edited byLam-Son Lê and Michel Toulouse. * Minh Thanh Chung [email protected] Nam Thoai [email protected] 1
Ho Chi Minh City University of Technology (HCMUT), 268 Ly Thuong Kiet Street, District 10, Ho Chi Minh City, Vietnam
2
Vietnam National University Ho Chi Minh City, Linh Trung Ward, Thu Duc District, Ho Chi Minh City, Vietnam
3
MNM-Team, Ludwig-Maximilians-Universitaet (LMU), Oettingenstraße 67, 80538 Munich, Germany
to close. As a result, educational data gathered from these systems expand more quickly in this day and age. Student’s performance and their behaviors can be better understand when these data are thoroughly examined. Therefore, these findings can help in identifying students’ risks to timely intervene, discover their hidden potentials, predict student’s performance in the next semester, etc. Based on the literature review in 2013 [8], we consider two main groups: “Student Modeling” and “Decision Support Systems” in terms of EDM. Some widely used methods are regression and classification for predictin
Data Loading...
