Predicting Student Degree Completion using Random Forest
Abstract
Recent reports indicate that 40 percent of freshman at four-year public colleges will not graduate. Further, the average completion rate for two-year community colleges is less than 40 percent. Therefore, increasing student retention rates in higher education is of great importance. Student retention is a measure of students' continued enrollment until graduation. To improve retention rates, colleges and universities require strategies for intentional advising to ensure that students are able to complete their majors in a timely manner. Currently, efforts have been made to adjust admission requirements; however, retention rates are still considered low and these strategies have reduced access to higher education for students from different economic sectors. Thus, institutions have recognized the need to understand the factors that impact retention to better focus their efforts. To this end, this research presents the application of random forests to predict degree completion within three years, which represents 150 percent time to completion, and identify the variables that impact student retention at a large community college in the Midwest. The random forest algorithm consists of bagging (combining) decision trees created randomly from the training sample, thus creating a 'forest . The model in this study was developed using data on 282 students with 14 variables. The variables included student details such as age, gender, degree, and college GPA. The model results, which include prediction and variable ranking, offer an important understanding about how to develop a more efficient and responsive system to support students.
Recommended Citation
T. A. Cardona et al., "Predicting Student Degree Completion using Random Forest," ASEE Annual Conference and Exposition, Conference Proceedings, article no. 1118, American Society of Engineering Education, Jun 2020.
Department(s)
Engineering Management and Systems Engineering
International Standard Serial Number (ISSN)
2153-5965
Document Type
Article - Conference proceedings
Document Version
Citation
File Type
text
Language(s)
English
Rights
© 2024 American Society of Engineering Education, All rights reserved.
Publication Date
22 Jun 2020
