Teaching Effectiveness of Higher Education Based on Big Data and Artificial Intelligence: A Novel Neutrosophic Multi-Criteria Decision Framework for Intelligent Educational Performance Evaluation

Abstract

The recent fast-paced digital revolution of education has led to an unprecedented amount of educational data being collected through learning management systems, online exams, student information systems, interactions in class and institutional analytics. Although the big data and artificial intelligence (AI) can help to assess teaching performance quite effectively, the current evaluation methods use mostly deterministic or fuzzy models, which cannot sufficiently reflect the uncertainty and inconsistencies of educational settings. Different opinions of experts, heterogeneous feedbacks, unobserved data and conflicting performance criteria often decrease the reliability and consistency of assessment of teaching effectiveness. In order to overcome these limitations, a new neutrosophic multi-criteria decision-making (MCDM) framework is introduced in this paper, which combines big data analytics and artificial intelligence for the evaluation of teaching effectiveness in higher education. This framework uses truth, indeterminacy and falsity membership degrees to model educational information, thus allowing one to evaluate simultaneously the certainty, ambiguity and contradictions within multiple criteria of teaching quality. Criterion weights are determined objectively based on characteristics of educational data, while the AI-based analytics help to identify important patterns of performance from large academic datasets. This model integrates the components via a well-constructed neutrosophic ranking process mathematically for obtaining trustworthy and understandable evaluations of teaching performance. The model has been formulated for the purpose of facilitating institutional decision-makers to identify their strengths in teaching, prioritize their professional development, and improve education quality via evidence-based decision-making processes. The proposed approach will be examined and compared with existing MCDM methods on a dataset regarding higher education teaching evaluation under the same experimental conditions. The sensitivity, robustness, and comparative studies will be carried out for investigating the consistency and feasibility of the proposed approach. The anticipated contribution is an intelligent educational evaluation framework that can successfully manage the uncertainty while taking advantage of big data and artificial intelligence. 

DOI: 10.5281/zenodo.22803608