Skip to main content

Table 2 Description of the variables in the assembled dataset

From: Modeling retroreflectivity degradation of pavement markings across the US with advanced machine learning algorithms

Variable types

Variables

Notations

Units

Descriptions

Numerical

Elapsed time

E

month

12 values (E0, E1, E2, E3, E11, E12, E15, E21, E24, E27, E33, and E36). E0 represents the month of installation, E1 represents 1 month since installation, E11 represents 11 months since installation, and similarly for other values

Measured Rs

MRS

mcd/m2/lux

12 values (MRS0 through MRS36, corresponding to E0 through E36)

Cumulative traffic level

TR

vehicles

12 values (TR0 through TR36, corresponding to E0 through E36)

Cumulative snowfall

SN

in

12 values (SN0 through SN36, corresponding to E0 through E36)

Marking thickness

TH

mil

172 unique values of marking thickness

Categorical

Surface type

S

-

2 unique categories (asphalt and concrete)

Types of marking material

T

-

7 unique categories (waterborne paint, thermoplastic, preformed thermoplastic, permanent polymeric tape, epoxy, polyurea, and MMA)

Marking color

C

-

2 unique categories (white and yellow)

Marking manufacturer

M

-

30 unique categories (M1 through M30)

Bead type of the first drop

b

-

7 unique categories (Type 1, Type 2, Type 3, Type 4, high-performance beads, wet reflective elements, and N/A. Here, N/A represents only one type of glass bead was applied as the second drop)

Bead type of the second drop

B

-

7 unique categories (Type 1, Type 2, Type 3, Type 4, high-performance beads, premium optics “Utah blend”, and N/A. Here, N/A indicates marking does not have dropped on glass beads)