Fig 1.
Research process.
Table 1.
Definition of near-crash events.
Fig 2.
Four typical directed acyclic graphs.
Table 2.
Conditional probability of node .
Fig 3.
Natural driving test route.
Table 3.
Data obtained from the natural driving test.
Table 4.
Statistics of near-crash events at different levels.
Table 5.
The average length of the road sections for each road type.
Table 6.
Example of matching risk points to road segment information.
Table 7.
The total risk score of each road segment.
Fig 4.
Risk points matched to road segments.
Table 8.
Number of different types of road sections with different risk levels.
Table 9.
Driving behavior Indicator system.
Table 10.
Value ranges corresponding to the high, medium, and low levels of different indicators.
Table 11.
Specifics of vehicle operation risk nodes.
Fig 5.
Structure of BN model for road segment risk identification.
Table 12.
Occurrence probability and conditional probability of observation nodes in different road types.
Fig 6.
Three types of BN models for road segment risk.
(a) BN model for risk of urban road sections. (b) BN model for risk of urban expressway sections. (c) BN model for risk of freeway sections].
Table 13.
Classification of road section operation risk levels.
Table 14.
Sensitivity analysis of driving risk on urban road sections.
Fig 7.
BN model when the risk index of urban road sections change.
(a) When the risk index of urban road sections decreased. (b) When the risk index of urban road sections increased.
Fig 8.
BN model when the risk index of urban expressway sections increased.
(a) When the risk index of urban expressway sections decreased. (b) When the risk index of urban expressway sections increased.
Fig 9.
BN model when the risk index of freeway sections increased.
(a) When the risk index of freeway sections decreased. (b) When the risk index of freeway sections increased.
Table 15.
Influence of Observed Node Changes on Model Results.