Professional Documents
Culture Documents
Miller
We are at critical junctures in several fields related to the broad scope of geo-
graphic analysis: (i) geographic information systems (GIS); (ii) transportation,
and; (iii) spatial analysis. By now, it is almost trite to point out the staggering
increase in digital information processing capabilities over the last three de-
cades and the resultant nse of GIS. As a suite of methods for capturing, storing,
analyzing, and communicating georeferenced information, GIS has spread to
many scientific fields concerned with understanding geographic phenomena. It
has also spread to application domains concerned with the efficient and equi-
table distribution of resources and the impact of resource use on the Earth's
environments.
The traditional geographic subfield of spatial analysis is witnessing a major
resurgence and enhancement due to GIS. For decades, spatial analysts tackled
fundamental problems associated with analyzing spatial data and modeling spatial
processes. As will be illustrated below, substantial progress has been achieved
The author thanks Tom Cova and three anonymous referees for helpful comments on an earlier
draft.
These require spatial sampling of representative locations that anchor travel such
as homes or workplaces. The latter includes roadside interviews, traffic coun-
ters, screen-line surveys (that is, data collection at locations crossing “natural”
boundaries such as rivers) and cordon surveys (that is, data collection at loca-
tions along the “artificial” boundary that demarcates the study area, although
sampling at internal cordons located at TAZ boundaries is also possible). These
require sampling of locations that intercept representative travel flows. These
data collection techniques are often complementary; for example, flow-based
surveys are necessarily brief and used to complement home or workplace-based
surveys (Ort6zar and Willumsen 1994).
Screen-line and cordon-based sampling can introduce measurement effects
similar to the MAUP. Only recording flows when some boundary is crossed
can lead to biased estimates of flow population characteristics. As the size of
each zone increases, there is a greater likelihood that shorter trips will be
missed since they do not cross some boundary (Kirby 1997; Ord and Cliff
1976; Rogerson 1990).
Spatial analysts have contributed to the problem of inferring flow population
characteristics when flows are sampled at boundaries. Ord and Cliff (1976) de-
velop methods for estimating population mean trip distances from boundary
sampled flows in the case of circular zones, randomly distributed populations,
and uniform, exponential, “half-normal’’ and normal density functions describ-
ing the distribution of trip distances in the study region. Rogerson (1990) esti-
mates mean migration distances for circular and square regions, randomly dis-
tributed and centralized populations, and exponentially distributed migration
distances. Rogerson (1990) also examines the influence of zonal shape on popu-
lation estimates, albeit in a limited manner. Kirby (1997) examines the effect of
screen-line and cordon-based sampling more explicitly. Similar to Ord and Cliff
(1976) and Rogerson (1990), Kirby (1997) examines a geometrically simple case
corresponding to a grid of screen-lines. Kirby (1997) also obtains results for the
case when flows follow “indirect” (that is, non-Euclidean) routes.
The methodologies developed by Ord and Cliff (1976), Rogerson (1990), and
Kirby (1997) require simple geometric structures and random distributions of
trip origins and destinations. These conditions are not likely to be met in prac-
tice. However, if we relax the requirement of analytical solutions, GIS can play
a central role in deriving numerical approximations. For example, a raster data
model can approximate the infinite locations within a zone with a finite lattice.
This can be used as a basis for approximating the intercept probabilities, mean
trip lengths and other population flow characteristics. These techniques could
exist as a component in a GIS software suite that allows the transportation ana-
lyst to adjust screen-line and cordon-based survey data before further analysis
and modeling.
Traffic counters are an inexpensive and unobtrusive way to collect flow data
within networks. Therefore, considerable attention has been directed in the trans-
portation modeling literature toward estimating or updating origin-destination
(0-D) matrices using these data. The basic problem is to estimate a new or update
an existing 0 - D matrix using flow data sampled at a subset of arcs in the transpor-
tation network. If the stud area is divided into M TAZs, the 0 - D matrix will
2
consist of either M 2 or ( M - M ) cells (the latter if intrazonal trips are ignored).
Theoretically, M 2 or ( M 2 - M ) independent and consistent traffic count observa-
tions are required to uniquely determine the 0 - D matrix. In practice, the num-
ber of sited traffic counters is often much lower. The looseness of the resulting
constraints means that many 0 - D matrices can be consistent with the sampled
traffic flows. The problem is to determine which 0 - D matrix should be selected.
Harvey]. Miller / 381
Often incorporated is a priori information such as a previous 0 - D matrix or a
postulated travel demand model (Ortlizar and Willumsen 1994).
Common estimation strategies for estimating 0 - D matrices from traffic
counts include generalized least squares (for example, Cascetta 1984), maxi-
mum likelihood (for example, Speiss 1987), Bayesian updating (for example,
Maher 1983) and entropy maximization (for example, Bell 1983; van Zuylen
and Willumsen 1980). Cascetta and Nguyen (1988) review estimation methods
within a unified formal framework while Ortlizar and Willumsen (1994) review
these methods from a pragmatic, problem-oriented perspective. Yang, Iida, and
Saski (1991) formulate an alternative error measure that provides improved
assessments of the reliability of the estimated 0 - D matrix.
Confounding factors when estimating 0 - D matrices from traffic counts are in-
consistent and interdependent count data (Ortlizar and Willumsen 1994). Mea-
surement errors from the counting devices and nonsimultaneous sampling can
lead to flow estimates that do not meet flow consistency requirements. These
consistency problems can be handled by treating the flow on.each network arc
as a random variable (Cascetta and Nguyen 1988).
Interdependent flow counts result from network flow structure being interde-
pendent, for example, the flow on a sampled link can result from flow on other
sampled links “upstream” on the shortest paths that use that link. This can re-
sult in redundant information in link counts. Bell (1983) discusses methods for
identifying linear dependencies in a link flow matrix; these include orthogonali-
zation and Gaussian elimination. However, these methods are computationally
intensive. Cascetta and Nguyen (1988) note that careful selection of sample
links can allow count data to be treated as independent. However, a link selec-
tion method is not evident in the transportation literature.
The problem of selecting locations for traffic counters to maximize indepen-
dence among flow counts can be viewed as a facility location problem where we
wish to minimize “cannibalizing” among sibling facilities (for example, stores
within the same retail chain). This is a well-studied principle in the location-
allocation literature (see Achabal, Gorr, and Mahajan 1982). It has also been
extended to situations involving demand related to traffic flows (for example,
Goodchild and Noronha 1987).
Hodgson (1990) and Hodgson, Rosing, and Zhang (1996) develop facility
location models that explicitly consider capturing or intercepting flows within a
transportation network. The $ow capturing location problem (Hodgson 1990) is
a “punitive” model oriented toward inspection programs that intercept vehicles
at any point in their journeys to deter violations such as driving under the influ-
ence of alcohol. In contrast, the preventive inspection model (Hodgson, Rosing,
and Zhang 1996) attempts to intercept vehicles as early as possible in their jour-
neys. This protects other network users from risks such as hazardous material
transport. Both models consider potential cannibalization of flows by facilities.
These models have obvious direct application in transportation management.
Their application or perhaps modification for traffic counter location and net-
work flow sampling is an open research question.
Intelligent transportation systems (ITS) create the possibility of complete or
near-complete recording of real-time traffic flows across an entire transportation
network (at least in those cities that can afford these systems). In this case, the
question of locating counting devices becomes moot, although facility location
methods can be used to select certain links if only samples of flows are required
for analysis. Measures of flow dependence among arcs are required if these
complete or near-complete data are used in models that assume independent
samples.
382 / Geographical Analysis
Black (1992) provides a method for assessing autocorrelation in network
flows. Blacks (1992) measures are special cases of the Moran’s I statistic that
assesses spatial autocorrelation among observations distributed in space (to be
discussed below). Spatial autocorrelation statistics such as Moran’s I require a
weight matrix reflecting the intensity of the spatial relationship between two
entities. These can be binary or real-valued and are typically based on nearest
neighbor relationships, distance, shared boundaries, etc. Black (1992) shows
that the weight matrix can be modified to reflect network structure. For exam-
ple, a binary weight can indicate if two arcs are incident to the same node or
real-valued weights can reflect shortest path distances within the network. Black
and Thomas (1998) apply this method to analyze accidents on motonvays in
Belgium. They also illustrate a method for determining the autocorrelation
source by decomposing the network and reanalyzing the autocorrelation mea-
sure independently for each component.
1.3 Spatial Dependency and Spatial Heterogeneity
1.3.1 Basic Issues. The vast majority of statistical theory and methods assumes
that observations are independent. This condition is rarely met when analyzing
spatial data. As Tobler (1970) points out through the oft-quoted First Law of
Geography, all things in space are related but things closer in space are more
related. Indeed, it is unclear why geographers and spatial analysts would be inter-
ested in any data or phenomena that are independent across space (Gould 1970;
quoted in Goodchild 1986). These dependencies imply that the results from
spatial analysis depend critically on the locations of the objects being analyzed.
Locational effects manifest themselves in two ways, namely, through spatial
dependence and spatial heterogeneity. The former effect is often referred to as
the spatial autocorrelation problem since it is a direct violation of the indepen-
dence assumption in mainstream statistical and econometric techniques. Spatial
heterogeneity occurs since every location has an intrinsic degree of uniqueness
due to its situation with respect to the rest of the spatial system. This results
in the estimated parameters of a spatial model being inadequate descriptors of
the process at any given location due to parameter drift across space (Anselin,
Dodson, and Hudak 1993). Consequently, ignoring spatial dependency and spa-
tial heterogeneity can invalidate the results of spatial data analysis (Anselin and
Griffith 1988).
Similar to the MAUP problem, recent years have witnessed resurgence of
attention on spatial dependency and spatial heterogeneity effects in spatial
data analysis. This is due to the increasing availability of digital geographic
databases and capabilities for handling geographic information. With respect
to spatial dependency, more sensitive measures of spatial autocorrelation are
emerging. An important breakthrough is recognition that spatial autocorrelation
effects are subject to spatial heterogeneity. This has resulted in measures that
capture local spatial autocorrelation effects; Anselin (1995) refers to these as
local indicators of spatial association or LISA statistics. Getis and Ord develop
the G statistic as a distance-based spatial autocorrelation measure that captures
spatial heterogeneity by comparing “local” weighted averages with global aver-
ages to identify spatial autocorrelation “hot spots” (Getis 1989, 1991; Getis and
Ord 1992, 1996; Ord and Getis 1995). Anselin (1995) derives local versions of
global spatial autocorrelation measures such as Moran’s I , the Geary statistic,
and the gamma statistic (see Constanzo, Hubert, and Golledge 1983; Hubert,
Golledge, and Constanzo 1981; Hubert et al. 1985). Bao and Henry (1996) for-
mulate generalized versions of the Moran and Geary LISA statistics and the G
statistic to account for irregular spatial units. Boots (1994) develops a technique
Harvey]. Miller f 383
forecasting, that is, explaining and predicting trip rates from origins in a study
area. Trip generations by TAZ or by household are regressed against explana-
tory variables such as land use (in the case of TAZs) or demographic factors
(in the case of households). Since these are geographic data, spatial dependence
and spatial hetereogeneity effects are likely. As discussed above, this creates
problems with parameter estimation, significance tests, and transferability to
other geographic contexts. Despite these potential problems, these issues are
not well recognized in the transportation modeling literature. This is despite
extensive discussion of other statistical problems such as multicollinearity and
nonlinearities. Only a small number of studies have examined trip generation
model transferability, typically reporting negative results (see Ortlizar and
Willumsen 1994).
GIS software provides a vehicle for capturing spatial dependence and spatial
heterogeneity effects in transportation modeling. GIS resolves two major
obstacles to widespread use of spatial statistics, namely, handling digital carto-
graphic structures and accessibility of spatial statistical methods. The availability
and easy handling of digital cartographic structures support the derivation of
weight matrices for spatial autocorrelation measures and spatial regression
models. Since CIS is increasingly used as a transportation analysis tool, integrat-
ing spatial statistics into GIS can put these methods in the hands of the trans-
portation analyst. Strategies for integrating GIS and statistical software has
received attention and several packages are now available (see Anselin 1992;
Anselin and Getis 1992; Anselin, Dodson, and Hudak 1993; Fotheringham and
Rogerson 1993).
A potential obstacle to wider use of spatial statistics is the computational
effort required for calculating these measures. The increasingly detailed and
global-scale nature of digital geographic data can overwhelm the capabilities
of serial computers regardless of their processing speeds. Parallel computing
architectures offer great promise in resolving this bottleneck, allowing spatial
statistics to scale upward and handle the very large geographic databases avail-
able now and in the future (Densham and Armstrong 1998; Griffith 1990). This
observation holds for other types of spatial analysis tasks: many spatial analytic
techniques can be decomposed into parallel computations either due to inher-
ent parallelism in the calculations or in the spatial data analyzed (Armstrong and
Densham 1992; Ding and Densham 1996; Xiong and Marble 1996). Armstrong
and Marciano (1995) and Armstrong, Pavlik, and Marciano (1994) report prom-
ising results with parallel implementations of the Getis-Ord G statistic. Experi-
ments with these types of parallel implementations should be conducted with
other spatial statistical and spatial analysis techniques, including the spatial re-
gression models discussed previously.
CI methods have been applied to modeling transportation and telecommuni-
cation flows. For example, CI methods such as Monte Carlo simulation, genetic
algorithms, and genetic programming can be used to “breed” spatial interaction
models from basic mathematical components in order to maximize goodness of
fit to a data set. This can generate unusual and surprising functional forms that
provide new insights into spatial interaction theory (see Openshaw 1988; Open-
shaw and Openshaw 1997, ch. 7). Diplock and Openshaw (1996) use genetic
algorithms and evolutionary strategies (a related, more flexible but computa-
tional intensive technique) to calibrate spatial interaction models with func-
tional forms that create difficulties with traditional calibration methods. A major
advantage of these approaches is their resilience and robustness: they can handle
parameter and variable values that result in overflow, underflow, and NaN errors
in traditional SI calibration techniques such as ordinary least squares (OLS) and
Harvey]. Miller / 387
MLE. CI-based calibration may require greater computational expense than
traditional methods, but as mentioned previously this is a concern of the past
rather than an issue of the future.
Artificial neural networks (ANNs) are alternatives to traditional travel demand
and spatial interaction models. ANNs are computational associative memory
structures inspired by the emergent properties that result from connectivity
among simple entities with binary states (that is, neurons) in biological neural
networks. Similar to other CI methods, ANNs are robust and fault tolerant, par-
ticularly with respect to noisy data. Like biological neural networks, ANNs are
especially adept at pattern recognition. ANNs learn from experience and “feel”
their way to a configuration of connectivity weights that associate inputs pat-
terns with output classifications to a high degree of accuracy (Openshaw and
Openshaw 1997; Peterson and Soderberg 1993). Because of their adeptness at
pattern recognition, ANNs are being used increasingly in classification and pre-
diction problems in geography (for example, Fischer et al. 1997; Foody 1995;
Gong, Pu, and Chen 1996; Gopal and Scuderi 1995; Lloyd 1994; Lloyd and
Carbone 1995). Other applications include multiattribute locational decision-
making (for example, Smith, Krishnamoorty, and Palaniswami 1996; Wang
1994; Zhou and Civco 1996) and theoretical explorations of the evolution of
urban structure (for example, White (1989)).
ANNs can be used for pattern recognition and related problems in transpor-
tation analysis (see Dougherty 1995). For example, ANNs have been used to
detect and predict vehicular accidents (Mussone, Rinelli, and Reitani 1996;
Ritchie and Cheu 1993), mode choice (Raju, Sikdar, and Dhingra 1996) and
transport and telecommunication flows (Black 1995; Fischer and Gopal 1994;
Fischer and Leung 1998; Gopal and Fischer 1996; Nijkamp, Reggiani, and
Tritapepe 1996). The advantages and disadvantages of ANNs in transportation
depend on the particular mode of analysis. ANNs are effective in exploratory
analysis when there is a large quantity of available data but no appropriate sta-
tistical theory to support analysis. ANNs are helpful when the data are incom-
plete, noisy, or fuzzy. ANNs are also effective in transportation forecasting, par-
ticularly for phenomena where behavioral theories at the individual level cannot
capture emergent macro properties (for example, dynamic propagation of con-
gested flows in transportation networks). However, a potential problem in this
case is ouerfitting, that is, training the A N N to reproduce a given dataset so well
that it cannot be transferred or generalized. Strategies for avoiding overfitting
include holding back part of the data set for validation in a manner similar to
good practice in traditional statistical estimation.
ANNs seem weakest in the explanato y mode of transportation analysis.
ANNs do not offer the theoretical insights available through models that are
well grounded in economic or behavioral theory. The connectivity weights in
a trained ANN do not have apparent economic or behavioral interpretations
that can aid in understanding the phenomenon under investigation (Nijkamp,
Reggiani, and Tritapepe 1996). However, under certain conditions a neural net-
work can be viewed as a weighted rule-based system whose outputs can be
interpreted as conditional probabilities. This can provide a theoretical frame-
work for interpreting the derived weights (see Buntine 1996).
Exploratory analysis and especially forecasting are increasingly important
issues in applied transportation analysis. In this respect, ANNs can be superior
to traditional modeling methods. Fischer and Gopal (1994) and Gopal and
Fischer (1996) report superior fit for ANNs under different learning regimes
relative to a spatial interaction model. However, the spatial interaction model
used as a benchmark is unconstrained: this typically produces inferior quality
388 / Geographical Analysis
results relative to singly or doubly constrained versions. Also, the spatial inter-
action calibration procedure used OLS rather than MLE: this typically leads
to underprediction of large flows and overprediction of small flows [see Fother-
ingham and O’Kelly (1989); Gopal and Fischer (1996) recognize this potential
problem]. Black (1995) reports better fit for an ANN applied to commodity flow
data relative to a doubly constrained spatial interaction model. Fischer and Leung
(1998) use genetic algorithms to search for the optimal weighted connectivity
pattern in a telecommunication-flow-predicting ANN. Results indicate that the
hybrid approach can automatically generate an optimal ANN topology, although
computational costs are higher than a pure ANN strategy. Nijkamp, Reggiani,
and Tritapepe (1996) are optimistic but cautious, noting that of the three
A N N architectures examined in their analysis, only one slightly outperformed
a logit model. Clearly required is more comparative research, particularly with
respect to different database sizes and data quality levels.
1.4 Alternative Representations of Geographic Environments
1.4.1 Basic Issues. Interaction and location are two sides of the same coin (Isard
1960). When analyzing spatial interaction, we often emphasize the influence of
location on interaction, for example, the relative propensity or attractiveness of
locations for generating or capturing flows. Frequently neglected are influences
in the other direction, that is, the influence of interaction on location, particu-
larly relative location as experienced or perceived by individuals. Understanding
experienced or perceived relative spaces can provide powerful insights into the
varying impact of transportation systems as well as the task environment facing
individuals when making travel and location decisions.
In addition to indirect impacts on land use, transportation and telecommu-
nication facilities can “warp” geographic space by changing the relationship
between time and space. Improvements in transportation technologies allow
individuals to trade less time for more space. This can result in a collapse of
geography or time-space convergence among certain locations (Janelle 1968,
1969). The close relationship between space-time accessibility and travel
implies that these locations are likely to experience increased interaction
(Kwan 1998; Thill and Horowitz 1997a, 1997b).
Levels of knowledge about the urban environment are also closely related to
interaction. An individual’s regular travel activity patterns provide the basis for
information gathering about the urban environment. Consequently, an individ-
ual’s spatial knowledge vanes widely across locations (Golledge 1978; Golledge
and Spector 1978; Stern and Leiser 1988). Knowledge of available opportunities
translates closely into destination and route choices, creating another positive
feedback loop similar to time-space convergence effects.
The interplay between travel behavior and urban space creates a dynamic
ecological structure to the city. Locations within an urban area have varying
mixes of employment, residential, retail, social, and recreational opportunities.
Individuals participate in different activities at different times depending on
diurnal, weekly, and longer activity cycles. This creates a complex but surpris-
ingly regular dynamic structure that cannot be adequately captured by a static,
two-dimensional representation of urban space (Goodchild, Klinkenberg, and
Janelle 1993; Janelle, Klinkenberg and Goodchild 1998).
These brief comments indicate the complex and dynamic interrelation-
ships between interaction and location. Different views of a travel environment
can be more relevant to individuals’ spatial behavior than the objective and
static view often used in transportation analysis. Understanding these alterna-
tive experience-based and cognitive-based views of an environment can provide
Harvey]. Miller / 389
insights into the spatial behavior of individuals as well as the location of eco-
nomic and social activities that respond to these behaviors.
1.4.2 Application: Visualizing the Efects of Transportation Systems. An
emerging research frontier in GISci is geographic visualization (GVIS). GVIS
can stimulate discovery of interesting and surprising relationships that escape
more rigid forms of analysis. GVIS is a superset of traditional cartographic
design. GVIS integrates principles of scientific visualization and cartographic
design within an interactive environment to reveal previously unknown relation-
ships and communicate these relationships to wider audiences (MacEachren
1995; MacEachren and Kraak 1997). The computational tools and design prin-
ciples of GVIS can be combined with models of relative space in spatial analysis
to develop powerful visualization techniques for transportation analysis.
Visualization techniques have been applied to transportation analysis, mostly
for analyzing flow matrices and flows within networks. For example, Liu (1995)
extends the “geographic brushing” technique for exploring multivariate geo-
graphic data (see Monmonier 1989) to develop a visualization system for explor-
ing migration flows. Dynamic flow visualization is becoming important as real-
time flow data gathered through ITS and dynamic transportation modeling
becomes increasingly prevalent. Ganter and Cashwell (1994) discuss some
design principles and a prototype system to support dynamic flow visualization.
Kwan and Hong (1998) and Miller (1991, 1999) use CIS to compute space-time
accessibility measures that capture the effect of the transportation networks on
accessibility. The resulting graphical depictions of spatial accessibility illustrate
the relationship between activity schedules, urban form, and accessibility and
provide striking depictions of varying accessibility in a study area. A worthwhile
topic for further investigation is visualization methods for exploring the poten-
tially enormous volume of data generated by real-time flow monitoring and in-
dividual accessibility measures applied at the urban scale.
The methods mentioned above attempt to visualize directly the performance
of transportation systems. Visualizing the impact of transportation on space is
indirect: this requires inferring the relative space implied by interactions or “dis-
tances” created by the transportation system. Several techniques exist for gen-
erating spaces given a (finite or infinite) set of objects and a relation describing
the similarity or dissimilarity among these objects. The resulting locational con-
figurations can be treated and analyzed as two-dimensional continuums similar
to the conventional map (Gatrell 1983; Tobler 1976a). These visualization tech-
niques have not diffused widely into spatial analysis or transportation analysis.
However, continuing development of GVIS principles and software can allow for
widespread use of these techniques, particularly for very large, high-resolution
geographic datasets.
In transportation analysis, the relations between locations can be travel cost,
travel time, interaction levels, or cognitive distances reported by individuals. If
the resulting matrix is symmetric we can use the techniques of multidimen-
sional scaling (MDS) and bidimensional regression to generate relative spaces
for visualization within GIs. MDS attempts to construct a space such that rela-
tive locations reflect as closely as possible the similarity/dissimilarity measures.
Although often used as a strategy for reducing multidimensional similarity/
dissimilarity relations, MDS is also effective at mapping the implicit functional
spaces resulting from travel times, cognitive distances, or spatial preferences.
Metric MDS uses real-valued dissimilarity measures while nonmetric MDS
uses ordinal information only (Gatrell 1983). Golledge and Spector (1978) and
Golledge, Rayner, and Rivizzigno (1982) use scaling techniques to generate rel-
ative spaces from distance estimates reported by individuals.
390 1 Geographical Analysis
work (Giuliano 1995). Instead, the transportation paradigm of the future will be
maximizing the efficiency of globally integrated multimodal transport systems.
Changing perspectives with respect to transportation solutions are occurring
at the same time that the tools available to researchers and practitioners are
changing radically. The explosion of available data and computational power
r
can not on1 provide new ways of analyzing transportation systems but even
new ways o thinking about these problems. The GISci community is providing
new methods for capturing, storing, rocessing, and communicating georefer-
K
enced information. In addition to hig lighting the relevance of spatial analysis
to CIS-T, this paper also suggests that a continuum exists between traditional
spatial analysis and the geocomputational tools emerging from GISci. A remark-
able synergy can only be achieved if the artificial boundaries among the trans-
portation, spatial analysis, and GISci communities continue to collapse.
LITERATURE CITED
Achabal, D. D., W. L. Gorr, and V. Mahajan (1982).“MULTILOC: A Multiple Store Location Decision
Model.”]mmal of Retailing 58(2), 5-25.
Amrhein, C. G. (1995). “Searching for the Elusive Aggregation Effect: Evidence from Statistical Simu-
lations.” Enoironment and Planning A 27, 105-19.
Amrhein, C. G., and R. Flowerdew (1992). “The Effect of Data Aggregation on a Poisson Regression
Model of Canadian Migration.” Enoironment and Planning A 24, 1381-91.
Angel, S., and G. M. Hyman (1976). Urban Fields: A Geometry of Mouemat for Regional Science.
London: Pion.
Anselin, L. (1988a).Spatial Econometrics: Methods and Models. Dordrecht: Kluwer Academic.
(1988b).“Lagrangian Multiplier Test Diagnostics for Spatial Dependency and Spatial Heteroge-
neity.” Geographicol Analysis 20, 1-17.
(1992).Spacestat: A Program for the Analysis of Spatial Data. National Center for Geographic
Information and Analysis, University of California, Santa Barbara.
(1993). “Discrete Space Autoregressive Models.” In Environmental Modeling with CIS, edited
by M. F. Goodchild, B. 0. Parks, and L. T. Steyaert, pp. 454-69. New York: Oxford university Press.
(1995).“Local Indicators of Spatial Association-LISA.” Geogmphicol Analysis 27,93-115.
Anselin, L., R. F. Dodson, and S . Hudak (1993). “Linking GIS and Spatial Data Analysis in Practice.”
Geographical Systems 1,3-23.
Anselin, L., and A. Getis (1992). “Spatial Statistical Analysis and Geographic Information Systems.”
A n d of Regional Science 26,19-33.
Anselin, L., and D. A. Griffith (1988).“Do Spatial Effects Really Matter in Regression Analysis? Papers
of the Regional science Association 65.11-34.
Anselin, L., and S. Rey (1991). “Properties of Tests for Spatial Dependence in Linear Regression
Models.” Ceographicol Analysis 23, 112-31.
Arbia, G., R. Benedetti, and G. Espa (1996).“Effects of MAUP on Image Classification.” Ceographicol
Systems 3, 123-41.
Arlinghaus, S . L., an&J. D. Nystuen (1990).“Geometry of Boundary Exchanges.” Geographicol Review
80.21-31.
Armstrong, M. P., and P. J. Densham (1992).“Domain Decomposition for Parallel Processing of Spatial
Problems.” Computers, Environment and Urban Systems 16,497-513.
Armstrong, M. P., and R. Marciano (1995).“Massively Parallel Processing of Spatial Statistics.” Znterna-
tional Jmmal of Geographical Systems 9, 169-89.
Armstrong, M. P., C. E. Pavlik, and R. Marciano (1994).“Experiments in the Measurement of Spatial
Association Using a Pardel Supercomputer.” Geographicol Systems 1,267-88.
Bao, S., and M. Henry (1996). “Heterogeneity Issues in Local Measurements of Spatd Association.”
Geographicol Systems 3, 1-13.
Batty, M. (1976).“Entropy in Spatial Aggregation.” Geographical Analysis 8, 1-31.
Batty, M. and P. K. Sikdar (1982a).“Spatial Aggregation in Spatial Interaction Models. 1. An Informa-
tion Theoretic Framework.” Enoironment and Planning A 14, 377-405.
(1982b). “Spatial Aggregation in Spatial Interaction Models. 2. One-Dimensional Population
Density Models,” Enuironment and Planning A 14, 525-53.
392 / Geographical Analysis
-(1982~).“Spatial Aggregation in Spatial Interaction Models. 3. Two-Dimensional Trip Distribu-
tion and Location Models.” Environment and Planning A 14, 629-58.
(1982d). “Spatial Aggregation in Spatial Interaction Models. 4. Generalisations and Large-Scale
Applications.” Environment and Planning A 14, 795-822.
Bearwood, J. E., and H. R. Kirby (1975). “Zone Definition and the Gravity Model: The Separability,
Excludability, and Compressibility Properties.” Transportation Research 9, 363-69.
Bell, M.G.H. (1983). “The Estimation of an Origin-Destination Matrix from Traffic Counts.” Transpor-
tation Science 17, 198-217.
Ben-Akiva, M., and S. R. Lerman (1985). Discrete Choice Analysis: Theory and Application to Travel
Demand. Cambridge, Mass: MIT Press.
Berry, B.J.L., and A. M. Baker (1968). “Geographic Sampling.” In Spatial Analysis: A Reader in Statis-
tical Geography, edited by B. J. L. Berry and D. F. Marble, pp. 91. Englewmd Cliffs, N.J.: Prentice-
Hall. 91-100.
Bivand, R. S. (1984). “Regression Modeling with Spatial Dependence: An Application of Some Class
Selection and Estimation Techniques.” Geographical Analysis 16, 25-37.
Black, W. R. (1992). “Network Autocorrelation in Transportation Network and Flow Systems.” Geo-
graphical Analysis 24, 207-22.
-(1995) “Spatial Interaction Modeling Using Artificial Neural Networks.” Journul of Transport
Geography 3, 159-66.
Black, W. R., and I. Thomas (1998). “Accidents on Belgium’s Motonvays: A Network Autocorrelation
Analysis.” ]ournu1 of Transport Geography 6, 23-31.
Boots, B. (1994). “Visualizing Spatial Autocorrelation in Point Data.” Geographical Systems 1, 255-66.
Boyce, D. E., Y.-F. Zhang, and M. R. Lupa (1994). “Introducing ‘Feedback into Four-Step Travel Fore-
casting Procedure Versus Equilibrium Solution of Combined Model.” Transportation Research Record
1443, 65-74.
Brunsdon, C., A. S. Fotheringham, and M. E. Charlton (1996). “Geographical Weighted Regression: A
Method for Exploring Spatial Nonstationarity.” Geographicd Analysis 28, 281-98.
Buntine, W. (1996). “Graphical Models for Discovering Knowledge.” In U. M. Fayyad, G. Piatetsky-
Shapiro, P. Smyth and R. Uthurusamy (1996) Adoances in Knowledge Discovery and Data Mining,
Cambridge, MA: MIT Press, 59-82.
Bunidge, P. (1980). “On the Cliff-Ord Test for Spatial Correlation.”]ournal of the Royal Statistical
Society B 42, 107-108.
Cascetta, E. (1984). “Estimation of Trip Matrices from Traffic Counts and Survey Data: A Generalized
Least Squares Estimator.” Transportation Research B 18B, 289-99.
Cascetta, E., and S. Nguyen (1988). “A Unified Framework for Estimating or Updating Origin/
Destination Matrices from Traffic Counts.” Transportation Research B 22B, 437-55.
Casetti, E. (1972). “Generating Models by the Expansion Method: Applications to Geographical Re-
search.” Geographical Analysis 4, 81-91.
Cervero, R. (1986). Suburban Gridlock. New Brunswick, N.J.: Center for Urban Policy Research.
Chou, Y-H. (1991). “Map Resolution and Spatial Autocorrelation.” Geographical
_ . Analysis 23, 228-46.
Clark, J. W. (1977).“Time-distance Transformationsof Transportation Networks.” Geographical Analysis
9, 195-205.
Clark, W. A. V., and K. L. Avery (1976). “The Effects of Data Aggregation in Statistical Analysis.” Geo-
graphical Analysis 8, 428-38.
Cliff, A. D. and P. Haggett (1970). “On the Efficiency of Alternative Aggregations in Region-Building
Problems.” Environment and Planning A 2, 285-94.
Cliff, A. D., and J. K. Ord (1972). “Testing for Spatial Autocorrelation among Regression Residuals.”
Geographical Analysis 4, 267-84.
Constanzo, C. M., L. J. Hubert, and R. G. Golledge (1983). “A Higher Moment for Spatial Statistics.”
Geographical Analysis 15, 347-51.
Cressie, N. (1992). Statistics for Spatid Data. New York: John Wiley and Sons.
-(1995). “Bayesian Smoothing of Rates in Small Geographic Areas.” ]ournu1 of Regional Science
35,659-73.
-(1996). “Change of Support and the Modifiable Areal Unit Problem.” Geographical Systems 3,
159-80.
Csillag, F., M. Kehesz, and A. Kummert (1996). “Sampling and Mapping of Heterogeneous Surfaces:
Multi-Resolution Tiling Adjusted to Spatial Variability.” Znternutiond ]ournu1 of Geographical Znfor-
mution Systems 10, 851-75.
H a r v e y ] . Miller / 393
Current, J. R. and D. A. Schilling (1987). “Elimination of Source a and b Errors in p-median Location
Problems.” Geographical Analysis 19, 95-110.
(1990). “Analysis of Errors Due to Demand Data Aggregation in the Set Covering and Maximal
Covering Location Problem.” Geographical Analysis 22, 116-26.
Daganzo, C. F. (1980a). “An Equilibrium Algorithm for the Spatial Aggregation Problem of Tr&c
Assignment.” Transportation Research B 14B, 221-28.
(1980b). “Network Representation, Continuum Approximations, and a Solution to the Spatial
Aggregation Problem of T&c Assignment.” Transportation Research B 14B, 229-39.
Densham, P. J., and M. P. Armstrong (1998). “Spatial Analysis.” In Parallel Processing Algorithm for
GIs, edited by R. Healy, S. Dowers, B. Gittings and M. Mineter pp. 387-413. London: Taylor and
Francis.
Ding, C. (1998). “The GIs-based Human-Interactive TAZ Design Algorithm: Examining the Impacts of
Data Aggregation on Transportation-Planning Analysis.” Enoironment and Planning B: Planning and
Design 25, 601-16.
Ding, Y., and P. J. Densham (1996). “Spatial Strategies for Parallel Spatial Modeling.” International
]ournu1 of Geographical Information Systems 10, 669-98
Diplock, G., and S. Openshaw (1996). “Using Simple Genetic Algorithms to Calibrate Spatial Interaction
Models.” Geographical Analysis 38, 262-79.
Dorigo, G., and W. R. Tobler (1983). “Push-Pull Migration Laws.’’ Annals of the Association of American
- .
Geocraphers 73, 1-17.
Dougherty, M. (1995). “A Review of Neural Networks Applied to Transport.” Transportation Research
C 3,247-60.
Downs, A. (1992). Stuck in Trafic: Coping with Peak-Hour Tra@ Congestion. Washington, DC: The
Brookings Institute.
Evans, S. P. (1976). “Derivation and Analysis of Some Models for Combining Trip Distribution and
Assignment.” Transportation Research 10, 37-57.
Ferguson, M. R., and P. S. Kanaroglou (1998). “Representing the Shape and Orientation of Destinations
in Spatial Choice Models.” Geographical Analysis 30, 119-37.
Fischer, M. M. (1997). “From Conventional to CI-based Spatial Analysis.” Paper presented at the 36th
annual meeting of the Western Regional Science Association, Hawai’i, F e b r u q 23-27.
Fischer, M. M., and S. Gopal (1994). “Artificial Neural Networks: A New Approach to Modeling Inter-
regional Telecommunication Flows.”]oumal of Regional Science 34, 503-27.
Fischer, M. M., S. Gopal, P. Staufer, and K. Steinocher (1997). “Evaluation of Neural Network Pattern
Classifiers for a Remote Sensing Application.” Geographical Systems 4, 195-223.
Fischer, M. M., and Y. Leung (1998). “A Genetic Algorithms Based Evolutionary Computational Neural
Network for Modeling Spatial Interaction Data.” A n d of Regional Science 32,437-458.
Foody, G. M. (1995). “Land Cover Classificationby an Artificial Neural Network with Ancillq Informa-
tion.” lntemutional]ournal of Geographical Information Systems 9, 527-42.
Foster, S. A. (1991). “The Expansion Method: Implications for Geographic Research.” Professional
Geographer 131-42.
Fotheringham, A. S. (1981). “Spatial Structure and Distance-Decay Parameters.” Annals of the Associa-
tion of American Geographers 71, 425-36
-(1983). “A New Set of Spatial Interaction Models: The Theory of Competing Destinations.”
Enoironment and Planning A 15, 15-36.
(1984). “Spatial Flows and Spatial Patterns.” Enoironment and Planning A 16, 529-43.
(1986). “Modeling Hierarchical Destination Choice.” Enoironment and Planning A 18, 401-
18.
Fotheringham, A. S., M. Charlton, and C. Brunsdon (1996). “The Geography of Parameter Space: An
Investigation of Spatial Non-stationarity.” International ]mrnal of Geographical Information Systems
10,605-27.
-(1997). “Two Techniques for Exploring Non-stationarity in Geographical Data.” Geographical
System 4,59-82.
Fotheringham, A. S., and M. O’Kelly (1989). Spatial Interaction Models: Fonnuhtions and Applications.
Dordrecht: Kluwer Academic.
Fotheringham, A. S., and P. A. Rogerson (1993). “GIS and Spatial Analytical Problems.” International
Joumul of Geographical Information Systems 7, 3-19.
Fotheringham, A. S., and M. J. Webber (1980). “Spatial Structure and the Parameters of Spatid Inter-
action Models.” Geographical Analysis 12, 33-46.
394 Geographical Analysis
Fotheringham, A. S., and D.W.S. Wong (1991). “The Modifiable Areal Unit Problem in Multivariate
Statistical Analysis.” Enoironment and Planning A 23, 1025-44.
Francis, R. L., T. J., Lowe, G. Rushton, and M. B. Rayco (1999). “A Synthesis of Aggregation Methods
for Mulifacility Location Problems: Strategies for Containing Error.” Geographical A d y k 31, 67-
87.
Ganter, J. H., and J. W. Cashwell (1994). “Display Techniques for Dynamic Network Data in Transpor-
tation GIS.” CIS-T ’94 Proceedings, 42-53.
Gatrell, A. C. (1983). Distance and Space: A Geographical Perspectiwe. Oxford: Oxford University Press.
Gehlke, C. E., and K. Biehl (1934). “Certain Effects of Grouping upon the Size of the Correlation Co-
efficient in Census Tract Material.” Journal of the American Statistid hsociatwn S u p p h t 29,
169-70.
Getis, A. (1989). “A Spatial Association Model Approach to the Identification of Spatial Dependence.”
Geographical Analysis 21, 251-59.
-(1991). “Spatial Interaction and Spatical Autocorrelation: A Cross-product Approach.” Enoiron-
rnent and Planning A 23, 1269-77.
Getis, A,, and J. K. Ord (1992). “The Analysis of Spatial Association by Use of Distance Statistics.” Geo-
graphical Analysis 24, 189-206.
-(1996). “Local Spatial Statistics: An Overview.” In Spatial A d y s i s : Modelling in a GZS enoiron-
ment. Edited by P. Longley and M. Batty, pp. 261-77. Cambridge, UK: GeoInformation Inter-
national.
Giuliano, G. (1995). “Land Use Impacts of Transportation Investments: Highway and Transit.” In The
Geography of Urban Transportation, 2d ed., edited by S. Hanson, pp. 305-41. New York: Guilford.
Golledge, R. G. (1978). “Learning about Urban Environments,” In Timing Space and Spacing Time I :
Making Sense of Time, edited by T. Carlstein, D. N. Parkes, and N. J. Thrift, pp. 76-98. London:
Edward Arnold.
Golledge, R. G . , J. N. Rayner, and V. L. Rivizzigno (1982). “Comparing Objective and Cognitive Repre-
sentations of Environmental Cues.” In Proximity and Preference: Problems in the Multidimensional
Analysis of Large Data Sets, edited by R. G. Golledge and J. N. Rayner, pp. 233-66. Minneapolis,
Minn.: University of Minnesota Press.
Golledge, R. G., and A. N. Spector (1978). “Comprehending the Urban Environment: Theory and Prac-
tice.” Geographical Analysis 10, 403-26.
Golledge, R. G., and R. J. Stimson (1997). Spatial Behaoior: A Geographic Perspectiwe. New York:
Guilford.
Gong, P., R. Pu, and J. Chen (1996) “Mapping Ecological Land Systems and ClassificationUncertainties
from Digital Elevation and Forest-Cover Data Using Neural Networks.” Photogrammetric Engineer-
ing and Remote Sensing 62, 1249-60.
Goodchild, M. F. (1977). “An Evaluation of Lattice Solutions to the Problem of Corridor Location.” En-
oironment and Planning A 9, 727-738.
-(1979). “The Aggregation Problem in Location-Allocation.”Geographical Analysis 11, 240-255.
-(1986). Spatial Autocorrelation. Concepts and Techniques in Modem Geography Number 47.
Nonvich, U.K.: GeoBooks.
-(1992). “Geographical Information Science.” International Journal of Geographic$ Znformation
Systenls 6, 31-45.
Goodchild, M. F., B. Klinkenberg, and D. G. Janelle (1993). “A Factorial Models of Aggregate Spatio-
Temporal Behavior: Application to the Diurnal Cycle.” Geographical Analysis 25, 277-94.
Goodchild, M. F., and V. T. Noronha (1987). “Location-Allocationand Impulsive Shopping: The Case of
Gasoline Retailing.” In Spatial Analysis and Location-AUocation Models, edited by A. Ghosh and G.
Rushton, pp. 121-36. New York: Van Nostrand Reinhold.
Gopal, S., and M. M. Fischer (1996). “Learning in Single Hidden-Layer Feedforward Network Models:
Backpropagation in a Spatial Interaction Modeling Context.” Geographical Analysis 28, 38-55.
Gopal, S., and L. Scuderi (1995). “Application of Artificial Neural Networks in Climatology: A Case
Study of Sunspot Prediction and Solar Climate Trends.” Geographical Analysis 27, 42-59.
Gould, P. R. (1970). “Is statistix inferens the Geographical Name for a Wild Goose?” Economic Geogra-
phy 46, 439-48.
Griffith, D. A. (1982). “Geometry and Spatial Interaction.” Annals of the Association of American Geog-
raphers 72, 332-46.
-(1983). “The Boundary Value Problem in Spatial Statistical Analysis.”Journal of Region$ Science
23, 377-87.
H a r v e y ] . Miller / 395
(1985). “An Evaluation of Correction Techniques for Boundary Effects in Spatial Statistical Anal-
ysis: Contemporary Methods.” Geographical Analysis 17, 81-88.
(1990). “Supercomputing and Spatial Statistics: A Reconnaissance.” Professional Geographer 42,
481-92.
Griffith, D. A., and C. G. Amrhein (1983). “An Evaluation of Correction Techniques for Boundary
Effects in Spatial Statistical Analysis: Traditional Methods.” Geographical Anulysis 15,352-60.
Gunn, H. F., M. E. Ben-Akiva, and M. A. Bradley (1985). “Tests of the Scaling Approach to Transfer-
ring Disaggregate Travel Demand Models.” Transportation Research Rewrd 1037, 21-30.
Hanson, S. (1995). “Getting There: Urban Transportation in Context,” in The Geography Of Urban
Transportation, 2nd ed., edited by S. Hanson, pp. 3-25. New York: Guilford Press.
Haslett, J., G. Wills, and A. Unwin (1990). “SPIDER-An Interactive Statistical Tool for the Analysis of
Spatially Distributed Data.” Znternutionul ]ournul of Geographical Znformution Systems 3, 285-96.
Hillsman, E. L., and R. Rhoda (1978). “Errors in Measuring Distances from Populations to Service Cen-
ters.” Annuls of Regional Science 12(3) 74-88.
Hodgson, M. J. (1990). “A Flow-Capturing hation-Allocation Model.” Geographical A d y s i s 22,270-
79.
Hodgson, M. J., K. E. Rosing, and J. Zhang (1996). “Locating Vehicle Inspection Stations to Protect a
Transportation Network.” Geographicd Analysis 28,299-314.
Holt, D., D. G. Steel, M. Tranmer (1996). “Area Homogeneity and the Modifiable Areal Unit Problem.”
Geographical Systems 3, 181-200.
Holt, D., D. G. Steel, M. Tranmer, and N. Wrigley (1996). “Aggregation and Ecological Effects in
Geographically Based Data.” Geogmphicol A d y s i s 28, 244-61.
Horn, M.E.T. (1995). “Solution Techniques for Large Regional Partitioning Problems.” Geographical
Analysis 27,230-48
Hubert, L. J., R. G. Golledge, and C. H. Constanm (1981). “Generalized Procedures for Evaluating Spa-
tial Autocorrelation.” Geographical Analysis 13, 224-33.
Hubert, L. J., R. G. Golledge, C. H. Constanm, and N. Gale (1985). “Measuring Association between
Spatially Defined Variables: An Alternative Approach.” Geographicd Anulysis 17, 36-46.
Hunt, L. and B. Boots (1996). “MAUP Effects in the Principal Axis Factoring Techniques.” Geographi-
cal systems 3, 101-21.
Isard, W. (1960). Methods of Regional Anulysis: An Introduction to Regional Science. Cambridge, Mass.:
MIT Press.
Janelle, D. G. (1968). “Central Place Development in a Time-Space Framework.” Profeessid Geogra-
pher 20,5-10.
(1969). “Spatial Organization: A Model and Concept.” Annuls of the Association of American
Geographers 59,348-64.
Janelle, D. G., B. Klinkenberg, and M. F. Goodchild (1998). ‘The Temporal Ordering of Urban Space
and Daily Activity Patterns for Population Role Groups.” Geographic$ Systems 5, 117-37.
Jones, J. P. 111 and E Casetti (1992). Applications ofthe Expansion Method. London: Routledge.
Keane, M. (1975). “The Size of the Region-Building Problem.’’ Environment and Planning A 7,575-77.
Kirby, H. R. (1997). “Buffon’s Needle and the Probability of Intercepting Short-Distance Trips by Md-
tiple Screen-Line Surveys.” Geographical Analysis 29, 64-71.
Koppelman, F. S., G.-K. Kuah, and C. G. Wdmot (1985). “Transfer Model Updating with Disaggregate
Data.” Transportation Research Rewrd 1037, 102-107.
Koppelman, F. S., and C. G. Wilmot (1982). “Transferability of Disaggregate Choice Models.” Trans-
portation Research Rewrd 895, 18-24.
Koshizuka, T. and 0. Kurita (1991). “Approximate Formulas of Average Distances hociated with
Regions and Their Applications to Location Problems.” Mathemutical Programming 52,99-123.
Kuiper, H. (1986). Distribution of Distances in Ptzgeographicul S p m . Aldershot, U.K.: Gower.
Kuiper, J. H. and J.H.P. Paelinck (1982). “Frequency Distribution of Distances and Related Concepts.”
Geographical Analysis 14,253-59.
Kwan, M.-P. (1998). “Space-Time and Integral Measures of Accessibility: A Comparative Analysis Using
a Point-based Framework.” Geographical Analysis 30, 191-216.
Kwan, M.-P. and X.-D. Hong (1998). “Network-based Constraints-oriented Choice Set Formation Using
GIS.” Geographical Systems 5, 139-62.
Larson, R. C., and A. R. Odoni (1981). Urban Operations Research. Englewood Chffs, N.J.: Prentice-
Hall.
396 J Geographical Analysis
LeBlanc, L. J., E. K. Morlok, and W. P. Pierskalla (1975). “An Efficient Approach to Solving the Road
Network Equilibrium Tr&c Assignment Problem.” Transportation Research 9, 309-18.
Leung, Y. (1978). “On the Imprecision of Boundaries.” Geographical Analysis 19, 125-51.
Liu, L. (1995). “PPFLOW: An Interactive Visualization System for the Exploratory Analysis of Migration
Flows.” Geographic lnfonnation Sciences 1, 118-23.
Lloyd, R. (1994). “Learning Spatial Protowes.” Annals of the Association of Ainerican Geographers 84,
4 18-40.
Lloyd, R., and G. Carbone (1995). “Comparing Human and Neural Network Learning of Climate Cate-
gories.” Professional Geographer 47, 237-50.
Lolonis, P., and M. P. Armstrong (1993). “Location-Allocationas Decision Aids in Delineating Adminis-
trative Regions.” Computers, Environment and Urban Systems 17, 153-74.
MacEachren, A. M. (1995). How Maps Work: Representation, Visualimtion and Design. New York:
Guilford.
MacEachren, A. M., and M.-J. Kraak (1997). “Exploratory Cartographic Visualization: Advancing the
Agenda.” Computers and Geosciences 23, 335-43.
Maher, M. (1983). “Inferences on Trip Matrices from Observations on Link Volumes: A Bayesian Statis-
tical Approach.” Transportation Research B 17B, 435-47.
Malanson, G., and M. P. Armstrong (1997). ‘‘Issues in Spatial Representation: Effects of Number of
Cells and Between-Cell Step Size on Models of Environmental Processes.” Geographic and Enoiron-
mental Modelling 1, 47-64.
Martin, R. J. (1987). “Some Comments on Correction Techniques for Boundary Effects and Missing
Value Techniques.” Geographical Analysis 19, 273-82.
Miller, H. J. (1991). “Modeling Accessibility Using Space-Time Prism Concepts within Geographical
Information Systems.” International ]ourno1 of Geographical Information System 5, 287-301.
(1996). “GIS and Geometric Representation in Facility Location Problems.” International lour-
nal of Geographical Information Systems 10, 791-816
-(1997). Towards Consistent Traoel Demand Estitnation in Transportation Planning: A Guide to
the Theonj and Practice of Equilibrium Traoel Demand Modeling. Research Report, Bureau of Trans-
portation Statistics, US. Department of Transportation. Available at http://www.bts.gov/tmip/papers/
feedbacWmiller/toc.htm.
-(1999). “Measuring Space-Time Accessibility Benefits within Transportation Networks: Basic
Theory and Computational Procedures.” Geographical Analysis 31(April 1999), 187-212.
Miller, H. J., and J. D. Storm (1996).“Geographic Information System Design for Network Equilibrium-
based Travel Demand Models.” Transportation Research C 4C, 373-89.
Miller, J. R. (1998). “Spatial Aggregation and Regional Economic Forecasting.” Annals of Regional Sci-
ence 32, 253-66.
Miron, J. (1984). “Spatial Autocorrelation in Regression Analysis: A Beginner’s Guide.” In Spatial Statis-
tics and Models, edited by G. L. Gaile and C. J. Willmott, pp. 201-22. D. Reidel.
Mitchell, J.S.B., and C. H. Papadimitriou (1991). “The Weighted Region Problem: Finding Shortest
Paths through a Weighted Planar Subdivision.”J o u m l of the Association for Computing Machinerrj
38, 18-73.
Moellering, H., and W. Tobler (1972). “GeographicalVariances.” Geographical Analysis 4, 34-50.
Monmonier, M. (1989). “Geographic Brushing: Enhancing Exploratory Analysis of the Scatterplot
Matrix.” Geographical Analysis 21, 81-84.
Mussone, L., S. Rinelli, and G. Reitani (1996). “Estimating the Accident Probability of a Vehicular Flow
by Means of an Artificial Neural Network.” Enoirontnent and Planning B: Planning and Design 23,
667-75.
Nakaya, T. (1997). “Statistical Inferences in Bidimensional Regression Models.” Geographical Analysis
29, 169-86.
Nijkamp, P., A. Reggiani, and T. Tritapepe (1996). “Modelling Interurban Transport Flows in Italy: A
Comparison between Neural Network Analysis and Logit Analysis.” Transportation Research C 4C,
323-38.
Okabe, A,, and H. J. Miller (1996). “Exact Computational Methods for Calculating Distances between
Objects in a Cartographic Database.” Cartography and Geographic Information Systetns 23, 180-
95.
Okabe, A,, and N. Tagashira (1996). “Spatial Aggregation Bias in a Regression Model Containing a Dis-
tance Variable.” Geographical Systems 3, 77-99.
HarveyJ. Miller 397
ONeill, W. A. (1991). “Developing Optimal Transportation Analysis Zones Using GIS.” ITE]ouml 61
(December), 33-36.
Openshaw, S. (1977a).“A Geographical Solution to Scale and Aggregation Problems in Region Building,
Partitioning, and Spatial Modeling.” Transactions of the Znstitute of British Geographers 2 (new
series), 459-72.
- (1977b).“Optimal Zoning Systems for Spatial Interaction Models.’’ Enoironment and Planning A
9, 169-84.
(1978). “An Optimal Zoning Approach to the Study of Spatially Aggregated Data.” In Spatial
Representation and Spatial Interaction, edited by I. Masser and P. J. B. Bown, pp. 93-113. Leiden:
Martinus Nijhoff.
(1988). “Building an Automated Modeling System to Explore a Universe of Spatial Interaction
Models.” Geographical Analysis 20, 31-46.
(1994). “Computational Human Geography: Towards a Research Agenda.” Enoironment and
Planning A 26,499-508.
(1996). “Developing GIS-relevant Zone-based Spatial Analysis Methods.” In Spatial Analysis:
Modelling in a GIS Enoironment, edited by P. Longley and M. Batty, pp. 55-73. Cambridge, U.K.:
GeoInformation International.
Openshaw, S., M. Charlton, and C. Wymer (1987). “A Mark I Geographical Analysis Machine for the
Automated Analysis of Point Pattern Data.” Zntmtional J o u m l of Geographical Information Sys-
tems 1,335-50.
Openshaw, S., and C. Openshaw (1997). ArtijiFcial Zntelligence in Geography. New York: John Wiley.
Openshaw, S., and L. Rao (1995). “Algorithms for Reengineering the 1991 Census Geography.” Enoi-
ronment and Planning A 27, 425-46.
Openshaw, S., and J. Schmidt (1996). “Parallel Simulated Annealing and Genetic Algorithms for Re-
engineering Zoning Systems.” Geographical Systems 3,201-20.
Openshaw, S., and P. J. Taylor (1979). “A Million or So Correlation Coefficients: Three Experiments on
the Modifiable Areal Unit Problem.” In Statistical Applications in the Spatial Sciences, edited hy N.
Wrigley, pp. 127-144. London: Pion.
Oppenheim, N. (1995). Urban Travel Demand Modeling: From lndioidual Choices to General Equilib-
rium. New York: John Wiley.
Ord, J. K., and A. D. Cliff (1976). “The Analysis of Commuting Patterns.” Enoironment and Planning A
8, 941-46.
Ord, J. K., and A. G e t i s (1995). “Local Spatial Autocorrelation Statistics: Distributional Issues and an
Application.” Geographical Afullysis 27, 286-306
Orhizar, J. de D., and L. G. Willumsen (1994). Modelling Transport, 2d ed. New York: John Wiley and
Sons.
Peterson, C., and B. Soderberg (1993). “Artificial Neural Networks.” In Modern Heuristic Techniques
for Combinatorial Problems, edited by C. R. Reeves, pp. 197-242. New York: John Wiley.
Plane, D. A,, and P. A. Rogerson (1994). The Geographical Analysis of Population with Applications to
Planning and Business. New York: John Wiley and Sons.
Raju, K. A., P. K. Sikdar, and S. L. Dhingra (1996). “Modelling Mode Choice by Means of an Artificial
Neural Network.” Enoironment and Planning B: Planning and Design 23, 677-83.
Ripley, B. D. (1979). “Tests of ‘Randomness’ for Spatial Point Patterns.”]ouml of the Royal Statistid
Society, Series B 41,368-74.
Ritchie, S . G., and R. L. Cheu (1993). “Simulation of Freeway Incident Detection Using Artificial Neu-
ral Networks.” Transportation Research C : 1, 203-17.
Robinson, W. S. (1950). “Ecological Correlation and the Behavior of Individuals.” American SoCMbgical
Reoiew 15,351-57.
Rodriguez-Bachiller, A. (1983). “Errors in the Measurement of Spatial Distances between Discrete
Regions.” Enuironment and Planning A 15, 781-99.
Rogerson, P. A. (1990). “Buffon’s Needle and the Estimation of Migration Distances.” Mathematical
Population Studies 2, 229-38.
Safwat, K.N.A., and T. L. Magnanti (1988). “A Combined Trip Generation, Trip Distribution, Modal
Split, and Trip Assignment Model.” Transportation Science 18, 14-30.
Sheffi, Y. (1985). Urban Transportation Networks: Equilibrium Analysis with Mathematical Program-
ming Methods. Englewood Cliffs, N. J.: Prentice-Hall.
Sheppard, E. S . (1979). “Gravity Parameter Estimation.” Geographical Analysis 11, 120-32.
398 1 Geographical Analysis
-(1984). “The Distance-Decay Gravity Model Debate.” In Spa& Statistics and Models, edited
by G. L. Gaile and C. J. Wilmott, pp. 366-88. D. Reidel.
Smith, K., M. Krishnamoorthy, and M. Palaniswami (1996). “Neural versus Traditional Approaches to
the Location of Interacting Hub Facilities.” Location Science 4, 155-71.
Smith, T. R., G. Peng, and P. Gahinet (1989). “Asynchronous, Iterative, and Parallel Procedures for
Solving the Weighted-Region Least Cost Path Problem.” Geographicnl Analysis 21, 147-66.
Speiss, H. (1987). “A Maximum Likelihood Model for Estimating Origin-Destination Matrices.” Trans-
portatiun Research B 21B, 395-412.
Stern, E. and D. Leiser (1988). Levels Of Spatial Knowledge and Urban Travel Modeling.” Geographi-
cal Analysis 20, 140-55.
ten Raa, T. (1983). “Distance Distributions.” Geographical Analysis 15, 164-69.
Thill, J.-C. and J. L. Horowih (1977a). “Modelling Nonwork Destination Choices with Choice Sets
Defined by Travel-Time Constraints.” In Recent Deuelopments in Spatial Analysis: Spatial Statistics,
Behaoioual Modelling and Computational Intelligence, edited by M. M. Fischer and A. Getis, pp.
186-208. Springer-Verlag.
-(1997b). “Travel Time Constraints on Destination-Choice Sets.” Geographicnl Analysis 108-
23.
Tobler, W. R. (1970). “A Computer Movie Simulating Urban Growth in the Detroit Region.” Economic
Geography 46,234-40.
(1976a). “The Geometry of Mental Maps.” In Spatial Choice and Spatial Behaoior: Geographic
Essays on the Analysis of Prefeences and Perceptions, edited by R. G. Golledge and G. Rushton,
pp. 69-81. Columbus: Ohio State University Press.
-(1976b).“Spatial Interaction Patterns.” J o u m l of Enoimnmental Systems 6, 271-301.
-(1978). “Migration Fields.” In Pupilation Mobility and Residential Change, edited by W. A. V.
Clark and E. G. Moore, pp. 215-32. Northwestern University Studies in Geography Number 25.
-(1987). “Experiments in Migration Mapping by Computer.’ American Cartographer 14, 155-63.
-(1994). “Bidimensional Regression.” Geographical Analysis 26, 187-212.
(1977). “Visualizing the Impact of Transportation on Spatial Relations.” Paper presented at the
36th Annual Meeting of the Western Regional Science Association, Big Island of Hawai’i, February
23-27.
Tobler, W. R., and S. Wmeburg (1971). “A Cappadocian Speculation.” Nature 231, 39-41.
van Bemmelen, J., W. Quak, M. van Hekken, and P. van Oosderom, (1993). “Vector vs. Raster-based
Algorithms for Cross-country Movement Planning.” Proceedings, Auto-Carto 11, 304-17.
van Zuylen, H. J., and L. G. Willumsen (1980). “The Most Likely Trip Matrix Estimated from Tr&c
Counts.” Transportation Research B 14B, 281-93.
Vaughn, R. (1984). “Approximate Formulas for Average Distances Associated with Zones.” Transports-
twn Science 18,231-44.
Wang, F. (1994). “The Use of Artificial Neural Networks in a Geographical Information System for
Agricultural Land-Suitability Assessment.” Enoironment and Planning A 26, 265-84.
Webber, M. J. (1980). “A Theoretical Analysis of Aggregation in Spatial Interaction Models.” Geograph-
ical Analysis 12, 129-41.
White, R. W. (1989). “The Artificial Intelligence of Urban Dynamics: Neural Network Modeling of
Urban Structure.’’ Papers of the Regional Science Association 67,43-53.
Wilkinson, G . G., F. Fierens, and I. Kanellopoulos (1995). “Integration of Neural and Statistical
Approaches in Spatial Data Classification.” Geographical Systems 2, 1-20.
Williams, P. A., and A. S. Fotheringham (1984). The Calibration of Spatial lntemction Models by Maxi-
mum Likelihood Estimation with Program SIMODEL. Geographic Monograph Series number 7,
Department of Geography, Indiana University.
Wilson, A. G. (1967). “A Statistical Theory of Spatial Distribution Models.” Transportation Research 1,
253-69.
-(1970). Entropy in Urban and Regional Modelling. London: Pion.
Wilson, R. M. (1990). “The Average Distance between Two Zones.” Geographical Analysis 22, 348-
51.
Wong, D., and C. Amrhein (1996). “Research on the MAUP: Old Wine in a New Bottle or Real Break-
through?” Geographical Systems 3, 73-76.
Wong, D.W.S., and A. S. Fotheringham (1990). “Urban Systems as Examples of Bounded Chaos:
Exploring the Relationship between Fractal Dimension, Rank-Size, and Rural-to-Urban Migration.”
Geografiska Annaler 72B, 89-99.
Harvey J. Miller f 399
Xiong, D., and D. F. Marble (1996). “Strategies for Real-Time Spatial Analysis Using Massively Parallel
SIMD Computers: An Application to Urban Traffic Flow Analysis.” l n t e m t i o d ] o u m l of Geo-
g r a p h i d Z n f m c i o n S y s t e m 10, 769-89.
Yang, H., Y. Iida, and T. Sasaki (1991). “An Analysis of the Reliability of an Origin-Destination Trip
Matrix Estimated from Traffic Counts.” Transportation Research B 25B, 351-63.
You, J., 2. NedovibBudic5, and T. J. Kim (1997a). “A GIS-based Tr&c Analysis Zone Design: Imple-
mentation and Evaluation.” Transportation Planning and Technology 21,69-91.
(199%). “A GIS-based Traffic Analysis Zone Design: Technique.” Transportation Planning and
Technology 21.45-68.
Zhou,J., and D. L. Civco (1996). “Using Genetic Learning Neural Networks for Spatial Decision Making
in GIS.” Photogrammetric Engineering and Remote Sensing 62, 1287-95.