F1@K decreased as K increased for all methods¶

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

Subclaims¶

q2 i? F1@K scores declined as K increased from 1 to 5 for every method evaluated, a pattern the authors attribute to the difficulty of predicting all tags correctly as the number of predicted tags grows. → Olatinwo 2025

Evidence¶

Olatinwo 2025¶

Olatinwo, S. O., & Demmans Epp, C. (2025). Predicting Tags for Learner Questions on Stack Overflow. International Journal of Artificial Intelligence in Education. https://doi.org/10.1007/s40593-024-00441-x

q2 · i? · design · r2

Cross-validation results across all six models show F1@K declining with K. The article calls this "a common phenomenon observed in multi-label tasks" and attributes it to the growing difficulty of predicting all tags correctly.

"It is worth noting that the F1@K score decreased as K increased for all methods. This is a common phenomenon observed in multi-label tasks (Xu et al., 2021; Zhou et al., 2019; He et al., 2022), where it can be challenging for a model to predict all tags correctly as the number of predicted tags increases, leading to a decrease in the F1@K score."

Discussion¶

Related Claims¶

• BERT-CBA outperformed baseline tag recommendation methods on Precision@K, Recall@K, and F1@K under cross-validation — related