Navigating Ethics in Digital Humanities: A Deep Dive into Decision Distribution within Higher Education at the University of Jordan

Abstract

This research aims to aid higher education institutions in making decisions that align with student needs and enhance their satisfaction. It considers decision presentation, timing of implementation, and communication of the benefits tied to educational quality improvements. To gauge student opinions, an online questionnaire research design was adopted, involving 3,000 male and female students from the University of Jordan. Findings indicated that students generally express dissatisfaction with higher education decisions and regulations due to unclear communication and limited implementation time.

For predicting educational quality outcomes, four machine learning algorithms were employed, each corresponding to four different higher education decisions. Notably, the Random Forest (RF) algorithm showcased superior performance. In the initial questionnaire, it achieved an accuracy of 97%, which slightly decreased to 92% in the second questionnaire due to the expanded dataset and varying factors affecting accuracy. The k-Nearest Neighbors (KNN) algorithm also yielded impressive results, achieving a remarkable 94% accuracy in the third questionnaire.

In the third questionnaire, the Decision Tree (DT) algorithm exhibited an accuracy of 85% in optimal scenarios. In contrast, the Convolutional Neural Network (CNN) algorithm, tailored for intricate tasks with numerous variables, consistently performed below expectations across all questionnaires. Its efficacy consistently lagged alternative algorithms, indicating a misalignment with the specific demands of its operational framework.