Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-0259
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-07-18T06:21:02.556165Z
status: published
lens: g_space
sector: education
headline: Den-SOFA: Dental Student Outcome Forecasting Assistant using explainable machine learning models
dek: Artificial intelligence (AI) and machine learning (ML) are increasingly being explored in dental education to support academic assessment and identify students at risk of poor performance. However, predictive modeling in this setting remains challenging because complex, nonlinear relationships among academic and demographic variables influence student achievement. Moreover, the value of such models lies not only in prediction but also in interpretability. This study evaluated Den-SOFA, an explainable ML framewor…
gain_title: Den-SOFA predicted pass/fail on restorative dentistry exit exams with AUC-ROC 0.906 and accuracy 0.86 using 26 academic and demographic variables from 96 students.
problem_title: (none)
trace_subject: (none)
gain_reading: Den-SOFA predicted pass/fail on restorative dentistry exit exams with AUC-ROC 0.906 and accuracy 0.86 using 26 academic and demographic variables from 96 students.
gain_evidence: For pass/fail prediction, the ANN model achieved the highest discrimination (AUC-ROC = 0.906; accuracy = 0.86; MCC = 0.71)
problem_reading: (none)
problem_evidence: (none)
quick_read: On 2026-07-16, a peer-reviewed study described Den-SOFA, an explainable machine learning framework tested as a proof-of-concept to forecast restorative dentistry exit examination outcomes from 26 academic and demographic variables for 96 dental students across five models.

The work matters because early identification of at-risk dental students could inform support, but the reported moderate-to-good performance was internal only, with no external validation and limited multiclass accuracy, leaving generalizability and operational utility unresolved.
limitation: Findings are limited to a single-institution dataset of 96 students and require external validation before operational use.
tag: Evidence-backed gain
key_points: Study analyzed 96 dental students with 26 academic and demographic variables including cumulative GPA and practical course grades. | Five models were compared for binary and multiclass prediction: logistic regression, random forest, XGBoost, CatBoost, and artificial neural networks. | Random forest showed comparable binary performance to ANN with AUC-ROC = 0.864 and accuracy = 0.86 but greater interpretability.
rundown: The analysis used cumulative GPA, theoretical and practical course grades, academic progression indicators, and demographic characteristics to predict both binary pass/fail and multiclass A-F grades on restorative dentistry exit examinations.

Performance was assessed using accuracy, F1-score, AUC-ROC, MCC, sensitivity, and specificity, with SHAP analysis used for interpretability; multiclass prediction reached only moderate performance with best AUC-ROC of 0.775 and demographic variables contributed minimally.
sources:
- peer_reviewed | BMC Medical Education | https://doi.org/10.1186/s12909-026-09949-3 | 2026-07-16
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

fc1ef65517a834911e544de97f8b49b202c6bc0344d80bedbfa5b83b541fbb26

previous

0000000000000000000000000000000000000000000000000000000000000000