Utilizing black-box visualization tools to interpret non-parametric real-time risk assessment models
- 25 August 2020
- journal article
- research article
- Published by Taylor & Francis in Transportmetrica A: Transport Science
- Vol. 17 (4) , 739-765
Abstract
This study bridges the gap between Real-Time Risk Assessment (RTRA) and its practical implications by following the post-hoc interpretability approach and utilizing black-box graphical tools for safety data visualization. The real-time traffic-related crash contributing factors were detected using the matched-case control design on 402-miles Interstate 80 in Wyoming. Four black-box visualization tools, including Partial Dependence Plot (PDP), Individual Conditional Expectation (ICE), centered ICE, and Accumulated Local Effect (ALE), were scrutinized to interpret the causal effect of these factors on crash probabilities. The results revealed that these techniques have many advantages, disadvantages, and unanswered questions that must be recognized by Active Traffic Management. PDPs must be accompanied by ICEs that explain the heterogeneity across observations. ALE is the most reliable technique in one-dimensional plots for highly correlated space of variables. However, there is a substantial distinction between PDP and ALE in two-dimensional plots that may make ALE an unreliable method.Keywords
Funding Information
- U.S. Department of Transportation (DTFH6116RA00007)
- Wyoming Department of Transportation (RS04218)
This publication has 40 references indexed in Scilit:
- Assessment of Interaction of Crash Occurrence, Mountainous Freeway Geometry, Real-Time Weather, and Traffic DataTransportation Research Record: Journal of the Transportation Research Board, 2012
- Estimation of Real-Time Crash RiskTransportation Research Record: Journal of the Transportation Research Board, 2011
- The statistical analysis of crash-frequency data: A review and assessment of methodological alternativesTransportation Research Part A: Policy and Practice, 2010
- Conditional variable importance for random forestsBMC Bioinformatics, 2008
- Crash Frequency Analysis with Generalized Additive ModelsTransportation Research Record: Journal of the Transportation Research Board, 2008
- Assessment of freeway traffic parameters leading to lane-change related collisionsAccident Analysis & Prevention, 2006
- Calibrating a Real-Time Traffic Crash-Prediction Model Using Archived Weather and ITS Traffic DataIEEE Transactions on Intelligent Transportation Systems, 2006
- Greedy function approximation: A gradient boosting machine.The Annals of Statistics, 2001
- Generalized Additive Models, Graphical Diagnostics, and Logistic RegressionGeographical Analysis, 1995
- Moving out of the Linear Rut: The Possibilities of Generalized Additive ModelsTransactions of the Institute of British Geographers, 1992