Integrating LECTOR topic preferences into at-risk prediction shows a tendency to improve performance, with significant AUC gains only in some cases¶

Claim · All claims
Evidence · 1 study (2 entries) · design r2 · q2
Currency · newest study 2025 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? In week 1 of course A-2019, out-of-class topic preferences achieved a significantly higher AUC than traditional representations, while in-class AUC and both F-score comparisons showed no significant difference. → Erwin Daniel López Zapata 2025 q2 i? Across the 60 prediction cases, results showed a tendency for topic preferences to improve predictive performance rather than a uniform significant advantage. → Erwin Daniel López Zapata 2025 (2)

Evidence¶

Erwin Daniel López Zapata 2025¶

Erwin Daniel López Zapata, Cheng Tang, Valdemar Švábenský, Fumiya Okubo, Atsushi Shimada. (2025). LECTOR: Summarizing E-book Reading Content for Personalized Student Support. Int J Artif Intell Educ 35. https://doi.org/10.1007/s40593-025-00478-6

q2 · i? · design · r2

Logistic Regression predictions on first-week course A-2019 data with 3-fold cross-validation: out-of-class AUC was 0.738±0.099 for T versus 0.613±0.178 for F (p-value 6.21E-6); "in the other cases there was no significant difference".

"out of class, topic preferences (T) achieved a significantly higher AUC than traditional representations (F), while in the other cases there was no significant difference ( p < 0.05)."

Erwin Daniel López Zapata 2025 (2)¶

Erwin Daniel López Zapata, Cheng Tang, Valdemar Švábenský, Fumiya Okubo, Atsushi Shimada. (2025). LECTOR: Summarizing E-book Reading Content for Personalized Student Support. Int J Artif Intell Educ 35. https://doi.org/10.1007/s40593-025-00478-6

q2 · i? · design · r2

Summary of the 60 prediction cases (about 14 datasets per course, courses A-2019 to A-2022, 600,712 logs from 218 students): the article characterizes the overall effect as "a tendency to improve the predictive performance by integrating LECTOR".

"The results showed a tendency to improve the predictive performance by integrating LECTOR."

Discussion¶

Related Claims¶

• Topic-preference features separate grade groups better than Reading Time in all analyzed cases — related
• Predictive models trained on one semester's offering of a course identified at-risk students in the subsequent semester with high prediction accuracy — related
• Random-forest models trained on a prior semester showed significantly lower AUC when tested on a new semester's data without retraining — related