EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Multimodal Performance Prediction in Online Learning Using Bi-LSTM and Mixture of Experts

Qikai Lin iD, Jingyu Liu, Yunhong Ding, Bo Qi iD

Sensors · 30.10.2026 · 26 (21) · 6874

DOI: 10.3390/s26216874 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 30.10.2026. Verfasst von Qikai Lin und 3 weiteren Autor:innen. Beteiligte Einrichtung: School of Computer Science and Information Engineering, Harbin Normal University, Harbin 150001, China. Hauptthemen laut MEDLINE: Learning. Die Arbeit nennt 56 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Heilongjiang Provincial Natural Science Foundation.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Educational records alone may not capture changes in students’ classroom affective states. This study proposes Bi-LSTM-SCPM, a student comprehensive performance model (SCPM) integrating bidirectional long short-term memory (Bi-LSTM) with mixture of experts (MoE) regression. The dataset comprised 14,645 facial images from 71 students across three stages of 12-week computer-assisted translation courses. ResNet50 achieved 93.8% accuracy in positive–negative facial state classification. Bi-LSTM integrated stage scores derived from instructor-designated key periods and Bayesian smoothing into a student-level affective score (ZHCJ). ZHCJ was combined with educational information and prerequisite course performance to predict final grades. Across 30 matched student-level splits, MoE achieved an R2 of 0.943 ± 0.002 and an MAE of 1.144 ± 0.128 points, compared with 0.927 ± 0.002 and 1.274 ± 0.149 points for XGBoost. Both differences remained significant after Holm adjustment (p < 0.001). ZHCJ also outperformed summary statistics and direct stage score inputs under both predictors. The setting ϖ = 2 and k = 20 performed best within the nine-setting grid. These findings support the predictive value of integrating temporal affective representations and educational information within the study cohort.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

LearningHumansLong Short Term MemoryBayes TheoremPredictive Learning ModelsStudents

Stichwörter der Autor:innen

academic performance predictioncamera-based sensingclassroom affective–behavioral analysismultimodal learning analyticsmultisource data fusiontemporal affective modeling

Autor:innen und Institutionen

  • School of Computer Science and Information Engineering, Harbin Normal University, Harbin 150001, China.
  • School of Astronautics, Harbin Institute of Technology, Harbin 150001, China.

ORCID-Profile: Qikai Lin, Bo Qi

Förderung

  • Heilongjiang Provincial Natural Science Foundation

Zitieren

APA

Lin, Q., Liu, J., Ding, Y., & Qi, B. (2026). Multimodal Performance Prediction in Online Learning Using Bi-LSTM and Mixture of Experts. Sensors, 26(21), 6874. https://doi.org/10.3390/s26216874

Vancouver

Lin Q, Liu J, Ding Y, Qi B. Multimodal Performance Prediction in Online Learning Using Bi-LSTM and Mixture of Experts. Sensors. 2026;26(21):6874. doi:10.3390/s26216874

BibTeX
@article{Lin2026,
  author = {Lin, Qikai and Liu, Jingyu and Ding, Yunhong and Qi, Bo},
  title = {Multimodal Performance Prediction in Online Learning Using Bi-LSTM and Mixture of Experts},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {21},
  pages = {6874},
  doi = {10.3390/s26216874}
}
Kostenlos starten →