Trip distribution forecasting with multilayer perceptron neural networks: A critical evaluation

Mozolin, Mikhail; Thill, Jean‐Claude; Usery, E. Lynn

doi:10.1016/s0191-2615(99)00014-4

Cited by 97 publications

(64 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In many cases, NNs outperformed the conventional models, leading to the conclusion that NNs may perform well enough to estimate spatial interaction flows in general. The only differentiating conclusion was presented by Mozolin, Thill, and Lynn Usery (2000) and Celik (2004). They concluded that NNs may perform better than conventional models for a base year matrix, but they fail to outperform conventional models for forecasting purposes.…”

Section: Modelling Trip Distribution With Nnsmentioning

confidence: 99%

“…Openshaw (1993) presented the potential use of NNs in spatial interaction modelling and Fischer and Gopal (1994) showed the applicability and predictive accuracy of NNs in modelling the distribution of interregional telecommunication flows. Many others have followed these pioneering works in trip distribution modelling: Black (1995) and Celik (2004) modelled commodity flows and Mozolin, Thill, and Lynn Usery (2000), Tillema, van Zuilekom, and van Maarseveen (2006) and Tapkın and Akyılmaz (2009) modelled intercity passenger flows with NNs.…”

Section: Modelling Trip Distribution With Nnsmentioning

confidence: 99%

See 1 more Smart Citation

Modelling trip distribution with fuzzy and genetic fuzzy systems

Kompil

Çelik

2013

Transportation Planning and Technology

View full text Add to dashboard Cite

This paper explores the potential capabilities of fuzzy and genetic fuzzy system approaches in urban trip distribution modelling with some new features. First, a simple fuzzy rule-based system (FRBS) and a novel genetic fuzzy rule-based system [GFRBS: a fuzzy system improved by a knowledge base learning process with genetic algorithms (GAs)] are designed to model intra-city passenger flows for Istanbul. Subsequently, their accuracy, applicability and generalizability characteristics are evaluated against the well-known gravity-and neural network (NN)-based trip distribution models. The overall results show that: traditional doubly constrained gravity models are still simple and efficient; NNs may not show expected performance when they are forced to satisfy trip constraints; simply-designed FRBSs, learning from observations and expertise, are both efficient and interpretable even if the data are large and noisy; and use of GAs in fuzzy rule-based learning considerably increases modelling performance, although it brings additional computation cost.

show abstract

Section: Modelling Trip Distribution With Nnsmentioning

confidence: 99%

Section: Modelling Trip Distribution With Nnsmentioning

confidence: 99%

Modelling trip distribution with fuzzy and genetic fuzzy systems

Kompil

Çelik

2013

Transportation Planning and Technology

View full text Add to dashboard Cite

show abstract

“…Using a logarithmic regression function as a benchmark model with three variables (two zonal magnitudes measured by gross regional products; and a friction variable), they conclude that the ANN outperforms the benchmark model. However, Mozolin et al (2000) note the underperformance of ANN prediction in comparison to the maximum likelihood doubly constrained gravity models in spatial interaction modeling of passenger flows among counties of the Atlanta Metropolitan Statistical Area in 1990. According to Mozolin et al, even if the literature reports calibration superiority of ANN in comparison to conventional models, an ex-post prediction of the conventional gravity model outperforms those of ANN.…”

Section: Literature Reviewmentioning

confidence: 99%

Modeling freight distribution using artificial neural networks

Çelik

2004

Journal of Transport Geography

View full text Add to dashboard Cite

“…The powerful and fast computing environment now available has brought many scholars to spatial interaction theory once again, either by utilizing evolutionary computation to breed novel forms of spatial interaction models (see Openshaw, 1988;Turton, Openshaw and Diplock, 1997) or applying neural network theory to spatial interaction, first proposed by Fischer and Gopal (1994) and later extended by many others [including Fischer and Leung, 1998;Bergkvist, 2000;Reggiani and Tritapepe, 2000;Mozolin, Thill and Usery, 2000;Fischer and Reismann, 2002a, b;Fischer, 2000Fischer, , 2002aFischer, Reismann and Hlavackova-Schindler, 2003].…”

Section: Introductionmentioning

confidence: 99%

Principles of Neural Spatial Interaction Modeling

Fischer

2009

Advances in Spatial Science

View full text Add to dashboard Cite

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may Principles of Neural SpatialInteraction Modelling MANFRED M. FISCHER Department of Economic Geography & Geoinformatics Vienna University of Economics and Business Administration AbstractNeural spatial interaction models represent the most recent innovation in the design of spatial interaction models. They are receiving increasing attention in recent years because of their powerful universal approximation properties. In essence they are devices for non-parametric statistical inferences, providing an elegant formalism for spatial interaction modelling. This contribution meets an urgent demand for methodological guidelines on how to develop robust applications that work from a statistical perspective. It introduces various components of a methodology for neural spatial interaction modelling that consists of a model specification framework to produce consistent estimators, a parameter estimation framework to compute parameter estimates that optimize an explicit fitness criterion and a framework to evaluate the model performance.JEL classification: C31, C45, R19

show abstract

Trip distribution forecasting with multilayer perceptron neural networks: A critical evaluation

Cited by 97 publications

References 14 publications

Modelling trip distribution with fuzzy and genetic fuzzy systems

Modelling trip distribution with fuzzy and genetic fuzzy systems

Modeling freight distribution using artificial neural networks

Principles of Neural Spatial Interaction Modeling

Contact Info

Product

Resources

About