Understanding speeding behavior from naturalistic driving data: Applying classification based association rule mining
DOI: 10.1016/j.aap.2020.105620
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This study investigates the complex associations between trip characteristics, driving behaviors, and roadway features that contribute to speeding, a significant factor in severe traffic crashes. Motivated by the need to understand speeding beyond single-factor analyses, the researchers aimed to identify combinations of variables that trigger specific speeding behaviors. The study focuses on two distinct perspectives: speeding duration (how long the speeding event lasts) and speeding pattern (the severity of the speed violation). By utilizing naturalistic driving data, the authors sought to provide transportation engineers and law enforcement with comprehensive insights to develop targeted countermeasures and calibrate simulation tools. The methodology employed Classification-Based Association (CBA) rule mining, a non-parametric algorithm that identifies frequent if-then associations without making subjective parametric assumptions. The analysis integrated two primary datasets: vehicle trajectory and surrounding vehicle data from the Safety Pilot Model Deployment (SPMD) program, and roadway inventory data from the Highway Performance Monitoring System (HPMS) for Michigan in 2016. Speeding events were defined as continuous driving periods exceeding the posted speed limit by more than 1 mph for longer than 30 seconds. These events were categorized into moderate or longer durations (based on a 2-minute threshold) and moderate or higher severity patterns (based on a 5 mph over-limit threshold). The final dataset comprised 2,812 speeding events, with variables including functional class, access control, shoulder width, median type, and prior speed loss. The results revealed distinct combinations of factors associated with different speeding behaviors. Longer speeding events (lasting more than 2 minutes) were highly associated with trips longer than 60 minutes and driving on roadways with higher functional classes, such as interstate highways. Conversely, moderate speeding durations were linked to shorter trips (less than 30 minutes), lower functional class roadways, and the absence of a median. Regarding speeding severity, higher speeding patterns (exceeding the limit by more than 5 mph) were triggered by combinations of lower functional class roadways, the presence of a median, and experienced congestion prior to the event. Moderate speeding patterns were associated with higher functional class roadways, shorter trips, and prior congestion. The CBA algorithm generated rule sets with high lift values, indicating strong interdependence between these antecedent features and the consequent speeding behaviors. The significance of this research lies in its demonstration that speeding behavior is likely triggered by composite effects of multiple factors rather than single variables. By distinguishing between duration and severity, the study offers a nuanced understanding of driver behavior that can inform more effective traffic safety strategies. The findings suggest that countermeasures should be tailored to specific contexts, such as addressing congestion-related speeding on lower-class roads or monitoring long-duration speeding on interstates. Additionally, the identified association rules provide valuable parameters for calibrating driver behavior in transportation simulation tools, enhancing the accuracy of predictive models used in traffic engineering and safety planning.
Key finding
Combinations of longer trips and higher functional class roadways are highly associated with longer speeding durations, whereas lower functional class roadways with prior congestion and medians trigger higher speeding patterns.
Methodology
naturalistic
Sample size: 2812
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via author_sweep_intake on 2026-05-27.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | author_sweep | — | — | 2 | 2026-05-27 |
| archive | success | canonical_url | — | — | 7 | 2026-06-06 |
| extract | success | cached | — | — | 3 | 2026-06-10 |
| clean | success | clean | — | — | 1 | 2026-06-04 |
| chunk | success | chunk | — | — | 1 | 2026-06-04 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-04 |
| enrich | skipped | — | — | — | 3 | 2026-06-04 |
| promote | success | — | — | — | 1 | 2026-06-04 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 2 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Empirical Findings: observational prevalence
- Methodological Resource: dataset resource