The effectiveness of adaptive learning platforms in higher education is shaped by multiple interacting factors

that involve subjective judgment and uncertainty. This study employs the Fuzzy Analytic Hierarchy Process

(Fuzzy AHP) to examine the relative importance of key evaluation criteria and sub-criteria based on the

perceptions of higher education users with prior experience in technology-enabled learning systems. Data were

collected through a structured survey of 150 respondents using fuzzy pairwise comparisons. The evaluation

framework was structured around four main criteria—Technological Factors, Content Quality, User Factors, and

Institutional Support—supported by twelve sub-criteria.

The findings show a clear and consistent emphasis on learner-oriented and practical considerations. Within the

technological dimension, system usability is ranked as the most influential factor, surpassing platform reliability

Overall, the results indicate that effective adaptive learning platforms are driven primarily by ease of use, high-

quality instructional content, and motivated learners, supported by responsive technical services. The study

provides a structured evaluation framework that can assist educators, institutions, and system designers in

making informed decisions regarding the development and adoption of learning platforms in higher education

contexts.

Keywords: Adaptive learning platforms; Fuzzy AHP; multi-criteria decision making; higher education; e-

learning evaluation

In the last decade, artificial intelligence (AI) has steadily shifted from an emerging technology to a central

efficient, and inclusive learning outcomes for diverse student populations (Tan et al., 2025) Adaptive learning is

not a new concept in educational research, but the integration of AI significantly enhances its capability by

enabling continuous analysis of large volumes of learner interaction data and adjusting content delivery

responsively. Early studies in e-learning have documented that AI and machine learning algorithms can

personalize content sequencing, optimize resource recommendations, and improve overall learner engagement

(Gligorea, 2023). Recent initiatives such as personalized adaptive programs in primary and secondary education

have already demonstrated meaningful gains. However, the adoption and evaluation of AI-based adaptive

learning platforms in higher education remain uneven and subject to contextual influences. Global comparisons

reveal that institutions in countries such as the United States, United Kingdom, Australia, China, and Singapore

are experimenting with diverse strategies for implementation, policy support, and technological investment, yet

Beyond measurable metrics like completion rates or test scores, educators and decision-makers must assess

factors such as personalization accuracy, pedagogical alignment, ethical use of data, technical integration,

scalability, user experience, cost, and institutional fit. These criteria are inherently subjective and often expressed

in linguistic terms that resist crisp numerical quantification. As a result, traditional decision-making methods

like simple weighted scoring or classical AHP (analytic hierarchy process) can fall short in capturing the

uncertainty and ambiguity inherent in stakeholder judgments. Multi-criteria decision analysis (MCDA) methods

have been increasingly used in educational research to address complex decision problems, but the need to

incorporate subjective judgments without forcing unwarranted precision calls for approaches that can model

fuzzy human reasoning (Alshakhatreh, 2024). The Fuzzy Analytic Hierarchy Process (Fuzzy AHP) has emerged

as a powerful extension of classical AHP precisely because it accommodates subjective judgments through fuzzy

In the context of adaptive learning technologies, applying Fuzzy AHP becomes particularly valuable because the

trade-offs between competing criteria cannot be resolved by objective measures alone. For example, a platform

with outstanding personalization capability may be costly and difficult to integrate with existing infrastructure;

another platform might be more affordable but offer weaker adaptive features. Reconciling such trade-offs

requires a decision framework that can handle both quantitative and qualitative considerations and synthesize

expert judgments into coherent priorities. Fuzzy AHP helps structure this complexity by decomposing the

evaluation into a hierarchy of criteria and sub-criteria uniquely tailored to the policy and pedagogical concerns

of higher education institutions.

reducing educational gaps (Ayeni et al., 2025). These findings reinforce the urgency of developing robust

evaluation frameworks that simultaneously respect technological performance and educational values.

From a global perspective, higher education is also navigating the broader implications of AI adoption beyond

adaptive learning alone. Institutions such as Purdue University and Northeastern University are embedding AI

competency and partnerships with AI system providers into curricula and campus strategies, reflecting a

recognition that AI is reshaping not just tools but the core of higher education . In this environment, adaptive

learning platforms represent one of the most visible intersections of AI, pedagogy, and institutional strategy —

and evaluating them systematically is essential for evidence-based decision making.

Yet few studies have integrated multi-criteria evaluation frameworks with fuzzy logic in the specific domain of

AI-based adaptive learning platforms in higher education. Most existing research either focuses on technology

development, algorithmic performance, or descriptive assessments of use cases. There remains a gap in literature

for systematic, comparative, and stakeholder-sensitive evaluation frameworks that are theoretically grounded,

methodologically rigorous, and practically relevant for institutional decision makers. This paper aims to fill that

systematically incorporated into the evaluation of adaptive learning platforms. It also offers a structured and

practical evaluation framework that higher education leaders, policymakers, and institutional decision makers

can employ to compare and select AI-based adaptive learning platforms in a transparent and methodologically

sound manner. Third, the study sheds light on the relative importance of pedagogical, technical, ethical, and

institutional factors that collectively shape the perceived value and effectiveness of adaptive learning systems.

By integrating expert input, linguistic weighting, and fuzzy logic aggregation, the proposed framework provides

a more comprehensive and realistic assessment of adaptive learning platforms than the evaluation approaches

commonly reported in existing literature.

The objectives of this study are threefold. First, it aims to identify and structure the key criteria and sub-criteria

The remainder of this paper is organized as follows. The next section reviews the relevant literature on AI-based

future research.

Technological factors (C1) constitute the core operational foundation of AI-based adaptive learning platforms

in higher education. Regardless of advances in adaptive algorithms or instructional design, platform

effectiveness ultimately depends on learners’ ability to access, navigate, and use the system without friction. In

adaptive environments, technical disruptions are particularly consequential because usability failures or system

interruptions can break personalized learning trajectories and undermine learners’ trust in the platform. System

usability, platform reliability, and internet accessibility therefore function as interdependent determinants of

whether adaptive learning technologies can deliver their intended pedagogical value.

likely to trust and accept personalization when the rationale behind content sequencing or feedback is apparent.

Platform reliability (SC12) complements usability by addressing the technical stability and dependability of the

system, including uptime, response speed, data accuracy, and consistent functioning of adaptive mechanisms.

Reliable platforms ensure uninterrupted access to learning activities, assessments, and feedback, which is

particularly critical in time-constrained higher education contexts. Evidence suggests that technical instability

diminishes learner confidence and reduces engagement, even when instructional design is pedagogically sound

(Bond et al., 2020). Internet accessibility (SC13) further extends the technological dimension by accounting for

Content quality(C2) represents the pedagogical substance of AI-based adaptive learning platforms and plays a

decisive role in shaping learning outcomes. While technological infrastructure enables system functionality, it is

the quality of instructional content that determines whether learning experiences are meaningful, engaging, and

aligned with academic objectives. Content relevance, interactivity, and multimedia support collectively define

the extent to which adaptive platforms can translate personalization into effective learning. Further, Content

relevance (SC21) refers to the alignment of learning materials with course objectives, disciplinary standards, and

individual learner needs. Adaptive systems enhance relevance by dynamically adjusting content sequencing,

difficulty levels, and learning pathways based on learner performance and preferences. Research consistently

learner progress. Multimedia support (SC23) refers to the integration of multiple instructional formats, including

text, video, audio, animations, and visual representations. When designed in line with cognitive principles,

multimedia content can enhance understanding and manage cognitive load by distributing information across

complementary channels (Mayer, 2020). Adaptive platforms can selectively deploy multimedia elements based

on learner data, reinforcing both personalization and instructional effectiveness. Collectively, these dimensions

position content quality as a dynamic, learner-centered construct within adaptive learning environments.

User factors (C3) encompass learner-related characteristics that shape how individuals interact with AI-based

adaptive learning platforms. These factors recognize that system effectiveness is influenced not only by

technological and content attributes but also by learners’ motivation, skills, and patterns of engagement. Learner

motivation (SC31) reflects the willingness to invest effort, persist through challenges, and pursue academic

platform use. Training availability (SC42) refers to structured programs that build users’ capacity to understand

and apply adaptive features effectively. Such training reduces resistance to technological change and promotes

pedagogically informed use of AI-based systems (Al-Azawei et al., 2019). Policy and management support

(SC43) captures institutional commitment through strategic planning, resource allocation, and governance

structures. Clear policies related to platform adoption, data governance, and pedagogical integration create an