EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Research on the english learning effect prediction model incorporating learning trajectory analysis

Ping Ma, Yongshun Tang, Di Liu

Data and Metadata · 17.05.2026 · 5 · 779

DOI: 10.56294/dm2026779 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Data and Metadata, online erschienen am 17.05.2026. Verfasst von Ping Ma und 2 weiteren Autor:innen. Die Arbeit nennt 27 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction

English learning impact prediction enhances personalized education by analyzing student behavior and performance. Existing research employs machine learning (ML) and Deep learning (DL) models for prediction, but many overlook dynamic learning trajectories and temporal variations. This constraint lowers prediction accuracy and undermines adaptive learning ideas.

Objective

This research proposes an English learning effect prediction model incorporating learning trajectory analysis to capture the dynamic evolution of learner performance and improve prediction.

Method

The English learning dataset was used to collect 7 540 samples from multiple learners via the Kaggle platform. Data preprocessing involves cleaning the dataset by eliminating missing values and min-max normalizing it before organizing it into sequential time-based learning records. Feature engineering is applied to extract trajectory-based indicators such as learning consistency, improvement rate, and performance fluctuation patterns to enhance model performance. The proposed Dynamic Wingsuit Flying Search-driven Vanilla Long Short-Term Memory (DWFS-Vanilla LSTM) model is used for English learning effect prediction model. A Vanilla LSTM model is used to identify temporal relationships in sequential learning behavior and forecast future learning performance. To improve prediction efficiency and model convergence, DWFS optimization is used to optimize the hyperparameters of the LSTM network.

Result

Experimental evaluation of a Python tool to demonstrate that the proposed trajectory-based Vanilla LSTM model significantly outperforms conventional deep learning models in terms of prediction accuracy (95,34 %), precision (96,25 %), recall (93,55 %), and F1-score (94,86 %). [user2.1][user2.2]

Conclusion

This research provides a robust data-driven model for intelligent English education systems and supports the development of adaptive learning strategies, early intervention mechanisms, and personalized instruction planning.

Zitieren

APA

Ma, P., Tang, Y., & Liu, D. (2026). Research on the english learning effect prediction model incorporating learning trajectory analysis. Data and Metadata, 5, 779. https://doi.org/10.56294/dm2026779

Vancouver

Ma P, Tang Y, Liu D. Research on the english learning effect prediction model incorporating learning trajectory analysis. Data and Metadata. 2026;5:779. doi:10.56294/dm2026779

BibTeX
@article{Ma2026,
  author = {Ma, Ping and Tang, Yongshun and Liu, Di},
  title = {Research on the english learning effect prediction model incorporating learning trajectory analysis},
  journal = {Data and Metadata},
  year = {2026},
  volume = {5},
  pages = {779},
  doi = {10.56294/dm2026779}
}
Kostenlos starten →