A Relative Behavior Standardization Framework for Cross-Learner Comparable Adaptive Learning Analytics in Metaverse Environments

Abstract

Metaverse-based learning environments generate rich and continuous behavioral data, including spatial movements, interaction activities, and temporal engagement patterns. While such data provide new opportunities for adaptive learning analytics, substantial inter-learner variability—stemming from differences in physical characteristics, interaction habits, and environmental conditions—limits the structural comparability and analytical reliability of existing approaches. Most prior studies rely on raw or lightly normalized behavioral signals, which may introduce magnitude bias and reduce cross-context reproducibility. This study proposes a Relative Behavior Standardization Framework designed to transform heterogeneous learner behavior data into structurally comparable and learner-invariant representations. The framework consists of four sequential stages: behavior decomposition, individual baseline modeling, relative behavior transformation, and structural alignment. By modeling behavioral signals relative to individual baselines rather than absolute magnitudes, the proposed approach mitigates inter-learner variability while preserving intrinsic spatiotemporal patterns. Unlike conventional normalization techniques that operate primarily at the numerical scale level, the proposed framework explicitly incorporates structural consistency across movement, interaction, and persistence dimensions. This design enhances cross-learner comparability, analytical stability, and reproducibility in adaptive learning systems. Conceptual validation and comparative analysis demonstrate the methodological advantages of the proposed approach over conventional non-standardized analytics. The proposed framework provides a stable analytical foundation for adaptive decision-making in metaverse-based education and contributes a structured methodology for behavior pattern standardization in immersive learning environments.