Unraveling Road Accident Risk Prediction Models With eXplainable AI
Bibliographic record
Abstract
Road traffic accidents pose a critical global challenge, leading to substantial fatalities and socioeconomic repercussions. Previous studies have examined intrinsic factors (e.g., driver age, vehicle speed) and extrinsic factors (e.g., weather, road conditions) that influence accident severity. However, the opacity of existing AI and machine learning (ML) prediction models hinders interpretability and limits their adoption in real-world applications. To address this gap, an explainable ML framework for accident severity prediction is proposed. Using diverse accident datasets, the predictive performance of multiple ML models is systematically evaluated to ensure adaptability across different geographic contexts. To enhance interpretability, explainable AI (XAI) techniques such as SHapley Additive exPlanations (SHAP), GeoShapely, and local interpretable model-agnostic explanations (LIME) are integrated to analyze key factors influencing accident severity. Additionally, an XAI evaluation framework is developed using rank consistency, rank alignment, importance stability, and fluctuation ratio to quantify interpretability. Our results indicate that refining ML models using top-ranked features from XAI improves prediction accuracy by 17.04%, 3.66%, and 5.06% for the three road accident datasets that we evaluated, namely the U.S., Ethiopia, and U.K. datasets, respectively. We also provide a decision flowchart to assist urban planners in choosing a suitable XAI approach for road accident severity prediction.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.019 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".