GPS Data and Car Drivers’ Parking Search Behavior in the City of Turnhout, Belgium - Geoinformatics for Intelligent Transportation - page 246

Fig. 5 Use of street segments for parking search

variables. The following hypotheses were de ned regarding the effects of the

included variables: a higher usage is expected for street segments.

Larger distance from the center ( (final destination);

￿

Short distance between street segment and parking facility;

￿

No presence of parking facilities;

￿

Presence of on-street parking;

￿

￿ Presence of shops or residences;

￿ Presence of two-way driving direction;

￿ Absence of dynamic parking guidance system;

￿ High parking tariff;

Secondary streets.

￿

first exploration of the relationship between the use of street segments and
the characteristics of street a basic linear regression model is estimated (Eq. 1 ).

For a

X

Ln ð F þ 0 5 Þ¼

b i X i

ð 1 Þ

:

i

where,
F

Number of times a street segment is used for searching

X i The i-th attribute of a street segment

The weight (parameter) of the i-th attribute.

β i

After evaluating several regression models, the best performing model only
includes four street segment attributes. The other attributes were removed from the
analyses because of a high correlation or the absence of a signi cant effect (Table 3 ).