Item Infomation
Title: | Usage of Machine Learning for Strategic Decision Making at Higher Educational Institutions |
Authors: | Nieto, Y. |
Participants: | GacÃa-DÃaz, V. Montenegro, C. González, C. C. Crespo, R. González |
Issue Date: | 2019 |
Publisher: | IEEE Xplore |
Series/Report no.: | IEEE Access, (2019), Vol 7, pp 75007-75017 |
Abstract: | Decisions made at the strategic level of Higher Educational Institutions (HEIs) affect policies, strategies, and actions that the institutions make as a whole. Decision's structures at HEIs are depicted in this paper and their effectiveness in supporting the institutions' governance. The disengagement of the stakeholders and the lack of using ef cient computational algorithms lead to 1) the decision process takes longer; 2) the ``whole picture'' is not involved along with all data necessary; and 3) small academic impact is produced by the decision, among others. Machine learning is an emerging eld of arti cial intelligence that using various algorithms analyzes information and provides a richer understanding of the data contained in a speci c context. Based on the author's previous works, we focus on supporting decision-making at a strategic level, being deans' concerns the preeminent mission to bolster. In this paper, three supervised classi cation algorithms are deployed to predict graduation rates from real data about undergraduate engineering students in South America. The analysis of receiver operating characteristic (ROC) curve and accuracy are executed as measures of effectiveness to compare and evaluate decision tree, logistic regression, and random forest, where this last one demonstrates the best outcomes. |
URI: | http://tailieuso.tlu.edu.vn/handle/DHTL/9925 |
Source: | http://doi.org/10.1109/ACCESS.2019.2919343 |
Appears in Collections: | Tài liệu hỗ trợ nghiên cứu khoa học |
ABSTRACTS VIEWS
21
VIEWS & DOWNLOAD
6
Files in This Item:
Bạn đọc là cán bộ, giáo viên, sinh viên của Trường Đại học Thuỷ Lợi cần đăng nhập để Xem trực tuyến/Tải về
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.