EUVIMEDEUVIMED entdecken
Crossref

Knowledge Graph–Empowered Learning Diagnosis and Adaptive Feedback in a Smart Course on English Pronunciation: Design, Implementation,and Effects

Liping An, Aihong Gao

Sci Online · 2026

DOI: 10.66106/soabcv.20260207 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sci Online, 2026. Verfasst von Liping An und 1 weiteren Autor:innen. Die Arbeit nennt 1 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Abstract:Diagnosing learners' pronunciation difficulties and delivering feedback that fits individual needs remain two persistent problems in English pronunciation instruction, particularly in large blended courses where teacher attention is scarce and learner differences are wide. This study reports the design, implementation, and evaluation of a knowledge graph–empowered diagnostic and adaptive feedback system embedded in a smart course on English phonetics and pronunciation for English-major undergraduates in China. Drawing on the theoretical resources of Learning Analytics, Cognitive Diagnostic Assessment, the Zone of Proximal Development,and feedback literacy research, we constructed a hierarchical pronunciation knowledge graph comprising segmental,suprasegmental, and connected-speech layers, and operationalised it as a diagnostic and feedback engine. The graph decomposes the domain into 127 atomic knowledge points organised through four relation types (prerequisite,inclusion, contrastive, and co-occurrence), and it drives two mechanisms: (1) a knowledge-state diagnosis that infers mastery probabilities from heterogeneous evidence, including automatic speech recognition (ASR) scores,oral test items, and self-reflections; and (2) an adaptive feedback mechanism that ranks and sequences feedback according to the learner's position in the graph rather than according to raw error frequency. A quasi-experimental design was adopted with 96 second-year English majors across two intact classes, one receiving graph-driven diagnosis and adaptive feedback and the other receiving conventional ASR feedback and whole-class review. Over a 16-week semester, the experimental group outperformed the comparison group on suprasegmental accuracy,comprehensibility ratings, and knowledge transfer tasks, while reporting higher feedback usefulness and stronger self-regulatory behaviour. Mediation analysis suggested that feedback internalisation partially mediated the treatment effect. The study also identified three tensions: over-diagnosis induced by dense evidence, learner avoidance of complex feedback, and the mismatch between graph granularity and instructional time. We argue that the value of a knowledge graph in language education lies less in visualisation than in its capacity to make instructional decisions accountable to a mapped domain, and we propose design principles for integrating such graphs into smart courses.

Zitieren

APA

An, L., & Gao, A. (2026). Knowledge Graph–Empowered Learning Diagnosis and Adaptive Feedback in a Smart Course on English Pronunciation: Design, Implementation,and Effects. Sci Online. https://doi.org/10.66106/soabcv.20260207

Vancouver

An L, Gao A. Knowledge Graph–Empowered Learning Diagnosis and Adaptive Feedback in a Smart Course on English Pronunciation: Design, Implementation,and Effects. Sci Online. 2026. doi:10.66106/soabcv.20260207

BibTeX
@article{An2026,
  author = {An, Liping and Gao, Aihong},
  title = {Knowledge Graph–Empowered Learning Diagnosis and Adaptive Feedback in a Smart Course on English Pronunciation: Design, Implementation,and Effects},
  journal = {Sci Online},
  year = {2026},
  doi = {10.66106/soabcv.20260207}
}
Kostenlos starten →