On 2019 Iowa crash narratives, the ALBERT model agreed with one expert on 73% of 100 disputed cases, but the evaluation is weakened by selection bias and missing methodological details.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Identification of Potentially Misclassified Crash Narratives using Machine Learning (ML) and Deep Learning (DL)
On 2019 Iowa crash narratives, the ALBERT model agreed with one expert on 73% of 100 disputed cases, but the evaluation is weakened by selection bias and missing methodological details.