Two-lane roads are categorized into 2 classes: I. Roads those users have favorite speed, like main roads and between cities roads.
Traffic density was categorized as no road within 200 m (reference category), road with < 2 cars/min, road with 2 5 cars/min, road with 6 10 cars/min, or road with > 10 cars/min.
We entered the annual mean PM2.5 concentration on a continuous scale and categorized distance to roads in four classes: ≤ 50 m, 51 100 m, 101 200, and > 200 m.
Based on the distribution of distance to road in this cohort and previous exposure studies showing exponential decay in exposures with increasing distance from a road, for our primary analyses we categorized distance to road as 0 50, 50 200, or ≥ 200 m (Adar and Kaufman 2007; Karner et al. 2010; Lipfert et al. 2006; Lipfert and Wyzga 2008; Sahlodin et al. 2007; Zhu et al. 2002).
Constrained based classification and spectral clustering are then applied on the selected features to categorize the road-users.
Data mining techniques [15] are further used to categorize the road accident locations and indentifying factors that affects accidents in those locations [16].
Digital maps of roads used for our analyses were obtained from the Ministry of Transportation, and were categorized into 4 types: national roads (R1), rural roads (R2), concession roads (R3), and local roads (R4).
The development of the road as a transport infrastructure took place primarily in the post-war period [6, 28] After World War II, with a few revisions, Japan's roads are categorized into the ones designated for general traffic (Douro-hou, "Road Act 1952" hereafter), and the other ones such as roads for forestry.
Roads are generally categorized as unpaved or paved with paved roads typically classified as flexible, rigid, or composite systems.
To conduct analysis of sensitivity of the results to the choice of cut-points, we also categorized distance to major roads as ≤ 150 m and > 150 m (Hoffmann et al. 2006; Venn et al. 2001).
