Institute of Transportation, MOTC

       In this study, we apply association rule mining to examine the relationships between red light running and other traffic violations on heavy vehicles. Besides, we take serious traffic violations and violations that require road safety lectures after convictions as the main research variables, to investigate the prevalent factors contributing to red light running violations among drivers of heavy vehicles and the subsequent violations that are more likely to occur after running red lights. To address the restriction of association rule algorithms in time-series analysis, we first aggregate the violation data and conduct the analysis with association rule algorithm. Results show that heavy vehicle drivers with the following violations are prone to running red lights: failing to follow instructions to enter the scales(§29-2IV), failing to fasten seatbelts(§31I), failing to fasten seatbelts on freeways or expressway(§31II), speeding, failing to maintain a safe driving distance, driving in lanes without following the instructions, etc.(§33I), violating regulation while making a turn or changing lanes(§48I), violating regulation while making a U-turn(§49), parallel parking(§56II), refuse to cooperate or stop for inspection, or flee the scene after being requested to stop by traffic officers(§60I), and failing to follow the directions or refuse to cooperate with traffic officers, etc.(§60II). Except for §31I, §31II, and §56II, all other aforementioned violations are serious traffic violations. This also implies that many serious traffic violations have relationships with running red lights. Besides, the results depict that heavy vehicle drivers tend to violate §60II after running red lights. Thus, it is possible to target potential red light running drivers through their violation records and impose prevention measures on them, as well as avoiding serious traffic violations(§60II) happening for those already running red lights.