TITLE:
Kriging Geostatistical Methods for Travel Mode Choice: A Spatial Data Analysis to Travel Demand Forecasting
AUTHORS:
Viviani Antunes Gomes, Cira Souza Pitombo, Samille Santos Rocha, Ana Rita Salgueiro
KEYWORDS:
Geostatistics, Kriging, Travel Mode Choice, Spatial Estimation
JOURNAL NAME:
Open Journal of Statistics,
Vol.6 No.3,
June
24,
2016
ABSTRACT: This paper aims to compare the results of
two techniques of Kriging (Ordinary Kriging and Indicator Kriging) that are
applied to estimate the Private Motorized (PM) travel mode use (car or
motorcycle) in several geographical coordinates of non-sampled values of the
concerning variable. The data used was from the Origin/Destination and Public
Transportation Opinion Survey, carried out in 2007/2008 at S?o Carlos (SP,
Brazil). The techniques were applied in the region with 110 sample points
(households). Initially, Decision Tree was applied to estimate the probability
of mode choice in surveyed households, thus determining the numeric variable to
be used in Ordinary Kriging. For application of Indicator Kriging it was used
the variable “main travel mode” in a discrete manner, where “1” represented the
use of PM travel mode and “0” characterized others travel modes. The results
obtained by the two spatial estimation techniques were similar (Kriging maps
and cross-validation procedure). However, the Indicator Kriging (KI) obtained
the highest number of hit rates. In addition, with the KI it was possible to
use the variable in its original form, avoiding error propagation. Finally, it
was concluded that spatial statistics was thriving in travel demand forecasting
issues, giving rise, for the both Kriging methods, to a travel mode choice
surface on a confirmatory way.