Symbolic Regression on Network Properties

Symbolic Regression on Network Properties

Märtens, M. (TU Delft Network Architectures and Services)
Kuipers, F.A. (TU Delft Embedded Systems)
Van Mieghem, P.F.A. (TU Delft Network Architectures and Services)

McDermott, James (editor)
Castelli, Mauro (editor)
Sekanina, Lukas (editor)
Wolf, Ina (editor)
García-Sánchez, Pablo (editor)

2017

Networks are continuously growing in complexity, which creates challenges for determining their most important characteristics. While analytical bounds are often too conservative, the computational effort of algorithmic approaches does not scale well with network size. This work uses Cartesian Genetic Programming for symbolic regression to evolve mathematical equations that relate network properties directly to the eigenvalues of network adjacency and Laplacian matrices. In particular, we show that these eigenvalues are powerful features to evolve approximate equations for the network diameter and the isoperimetric number, which are hard to compute algorithmically. Our experiments indicate a good performance of the evolved equations for several real-world networks and we demonstrate how the generalization power can be influenced by the selection of training networks and feature sets.

Cartesian genetic programming
Complex networks
Symbolic regression

http://resolver.tudelft.nl/uuid:afa57467-7ef9-41d5-a816-bf598c01cdd5

https://doi.org/10.1007/978-3-319-55696-3_9

Springer, Cham

978-3-319-55695-6

Genetic Programming: Proceedings - 20th European Conference, EuroGP 2017

EuroGP 2017, 2017-04-19 → 2017-04-21, Amsterdam, Netherlands

Lecture Notes in Computer Science, 0302-9743, 10196

Institutional Repository

conference paper

© 2017 M. Märtens, F.A. Kuipers, P.F.A. Van Mieghem