Prefix-based data augmentation for early prediction in educational sequences¶
Research Method · All research methods
Evidence · 3 claims (3 for) · 2 studies (1 associational, 1 design),q2· 0 of 2 report an effect size · 3 claims rest on one study
Description¶
A data augmentation technique in which, instead of sliding windows or cropping, "we generate all prefixes of each sequence (from length 1 to n), allowing the model to learn to make predictions at any point in a student's progression". Each student contributes twelve records, each masking all data after a given chapter, so models can be trained and evaluated on data available only through each chapter of the course.
Accounts¶
- Prefix-based data augmentation for early prediction in educational sequences: A data augmentation technique in which, instead of sliding windows or cropping, "we generate all prefixes of each sequence (from length 1 to n), allowing the model to learn to make predictions at any point in a student's progression". Each student contributes twelve records, each masking all data after a given chapter, so models can be trained and evaluated on data available only through each chapter of the course. (Alice Xu et al. (2026))
Claims¶
- Behavioral-only metrics from an instrumented online textbook support early prediction of final course grades in flipped statistics courses [+W]
- Predictive models built in a modelable ecosystem generalize across institutions [+W]
- Behavior-prediction accuracy rises with more past learning records, reaching 0.94 at 40 records [+W]
Related Research Methods¶
-
Key Sources¶
- Alice Xu, Icy (Yunyi) Zhang, Adam B. Blake and James W. Stigler. (2026). Modelability as a Strategy for Improving the Generalizability and Scalability of Predictive Models. Journal of Learning Analytics, 13(1). https://doi.org/10.18608/jla.2026.9099
Page metadata
| Field | Value |
|---|---|
type |
research-method |
id |
prefix-based-data-augmentation-for-early-prediction-in-educational-sequences |
title |
Prefix-based data augmentation for early prediction in educational sequences |
description |
A data augmentation technique in which, instead of sliding windows or cropping, "we generate all prefixes of each sequence (from length 1 to n), allowing the model to learn to make predictions at any point in a student's progression". |
status |
review |
moved_from |
element/prefix-based-data-augmentation-early-prediction |
generated |
by process:sweep-kinds on 2026-10-10 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — alice-xu-2026 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.