Neural Network Modelling of Constrained Spatial Interaction Flows



If out-of-sample [generalisation] performance is more important than fast learning, then
the modular product unit neural network exhibits clear superiority. As can be seen by
comparing the ARV-values and the SRMSE-values the modular product unit neural
network model ranks best, followed by the two-stage neural network approach and the
gravity model. The average generalisation performance, measured in terms of
ARV(
M3), is 0.2022 and, measured in terms of SRMSE(M3), 1.1017, compared to
0.2251 and 1.1614 in the case of the two-stage approach, and 0.2262 and 0.1658 in the
case of the gravity model15. This difference in performance between the modular
product unit neural network model and the benchmark models is statistically
significant16. If, however, the goal is to minimise execution time and a sacrifice in
generalisation accuracy is acceptable, then the standard origin constrained gravity
model is the method of choice. The gravity model outperforms the neural network
models in terms of execution time, the modular product unit network model by a factor
of 10 and the 2-stage neural network model by a factor of 102. But note that this is
mainly caused by two factors: first, that our implementations were done on a serial
platform even though the neural network models are parallelizeable, and, second, that
we implemented a rather time consuming termination criterion (
к = 40,000) to stop the
training process.

Table 3: Benchmark comparisons of the Modular Product Unit Neural Network 1CSL
with the Two-Stage Neural Network Approach Ω'''t and the Gravity
Model
τ g"av for Modelling Origin Constrained Spatial Interactions

Modular Product Unit

Neural Network

Two-Stage Neural

Network Approach

Origin Constrained

Gravity Model

In-Sample (Training)

Performance

ARV

SRMSE

0.2212 (0.0087)

1.2858 (0.0254)

0.2682 (0.0222)

1.4150 (0.0578)

0.2121 (0.0017)

1.2594 (0.0049)

Out-of-Sample (Testing)

Performance

ARV

SRMSE

0.2022 (0.0150)

1.1017 (0.0407)

0.2251 (0.0255)

1.1614 (0.0670)

0.2262 (0.0027)

1.1658 (0.0069)

Note: Figures represent averages taken over 60 simulations differing in the initial
parameter values randomly chosen from [-0.3, 0.3] (standard deviations in
brackets); the testing set consists of 248 patterns and the training set of 496 patterns.

28



More intriguing information

1. Reform of the EU Sugar Regime: Impacts on Sugar Production in Ireland
2. Education Responses to Climate Change and Quality: Two Parts of the Same Agenda?
3. The name is absent
4. On the origin of the cumulative semantic inhibition effect
5. Update to a program for saving a model fit as a dataset
6. Clinical Teaching and OSCE in Pediatrics
7. What Contribution Can Residential Field Courses Make to the Education of 11-14 Year-olds?
8. AGRICULTURAL TRADE IN THE URUGUAY ROUND: INTO FINAL BATTLE
9. CURRENT CHALLENGES FOR AGRICULTURAL POLICY
10. Review of “The Hesitant Hand: Taming Self-Interest in the History of Economic Ideas”