Model-family rankings depend on the prediction target: AdaBoost leads the five-level ordinal task while LightGBM holds the highest binary macro-F1¶
Claim · All claims
Evidence · 1 study · designr2·q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly
Subclaims¶
Reading the evidence codes
q is the design tier, i the impact, n the sample, and the kind with its r
the rigour judged within that kind: what each code means.
This claim: no standardized effect is recorded in observations/ for it yet, so no benchmark reading is possible.
q2 i? On the secondary binary target, LightGBM achieved the highest macro-F1 point estimate (58.9%±7.9), ahead of XGBoost (56.7%±5.1) and AdaBoost (54.3%±3.0), reversing the ordinal-task ordering. → Sidharth Anupkrishnan 2026
Evidence¶
Sidharth Anupkrishnan 2026¶
Sidharth Anupkrishnan, Itir Sayar, Jeongah Lee, Sri Harsha Musunuri, Guan-Ming Su, Madeline Endres, Ravi Karkar, Phuc Nguyen. (2026). E3Sense: Head-Confined Multimodal Sensing of Learner Engagement. arXiv. https://arxiv.org/abs/2609.26569 doi:10.48550/arxiv.2609.26569
q2 · i? · design · r2
Binary sensitivity analysis within the same five-fold participant-grouped RQ1 evaluation, using the fixed 166-dimensional representation. The study reports the LightGBM "macro-F1 point estimate at 58.9%±7.9", above the 49.0%±1.1 random-distribution binary baseline; values are descriptive fold means.
"LightGBM has the highest macro-F1 point estimate at 58.9%±7.9, followed by XGBoost at56.7%±5.1and AdaBoost at54.3%±3.0."
Discussion¶
Related Claims¶
- Fold-local top-20 feature selection had mixed effects across model families, supporting the complete 166-dimensional representation as the primary analysis — related
- A head-confined multimodal configuration (EEG, eye tracking, forehead EDA) predicts held-out learners' five-level engagement ratings at 75.0% balanced 1-off accuracy, 12.0 points above the mode baseline — related
- Supplying learners' own engagement-definition codes as context raised exploratory prediction from 64.6% to 71.5% balanced 1-off accuracy in an eighteen-participant cohort
- Predictive performance of sensor configurations does not change monotonically with the number of sensing streams; preferred configuration depends on target metric and sensing cost — a broader claim this one bears on
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
model-family-rankings-depend-on-target |
title |
Model-family rankings depend on the prediction target: AdaBoost leads the five-level ordinal task while LightGBM holds the highest binary macro-F1 |
description |
Model-family rankings depend on the prediction target: AdaBoost leads the five-level ordinal task while LightGBM holds the highest binary macro-F1 |
status |
review |
evidence_strength |
moderate |
newest_evidence |
2026 |
recheck_by |
2029 |
generated |
by process:wiki-ingest on 2026-10-10 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — sidharth-anupkrishnan-2026 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.