nach oben

Evolutionary Intelligence

Erschienen in:

01.06.2012 | Special Issue

XCSF with local deletion: preventing detrimental forgetting

verfasst von: Martin V. Butz, Olivier Sigaud

Erschienen in: Evolutionary Intelligence | Ausgabe 2/2012

Einloggen

Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.

search-config

KI-gestützte Suche

Aus

Abstract

The XCSF classifier system iteratively solves regression problems with a population of overlapping, local approximators. We show that problem solution stability and accuracy may be lost in particular settings—mainly due to XCSF’s global deletion. We introduce local deletion, which prevents these detrimental effects to large extents. We show experimentally that local deletion can prevent forgetting in various problems—particularly where the problem space is non-uniformly or non-independently sampled. While we use XCSF with hyper-ellipsoidal receptive fields and linear approximations herein, local deletion can be applied to any XCS version where locality can be similarly defined. For future work, we propose to apply XCSF with local deletion to unbalanced, non-uniformly distributed, locally sampled problems with complex manifold structures, within which varying target error values may be reached selectively.

Vorheriger Artikel Function approximation with LWPR and XCSF: a comparative study

Nächster Artikel On principal component analysis for high-dimensional XCSR

Sie haben noch keine Lizenz? Dann Informieren Sie sich jetzt über unsere Produkte:

Springer Professional "Wirtschaft+Technik"

Online-Abonnement

Mit Springer Professional "Wirtschaft+Technik" erhalten Sie Zugriff auf:

über 102.000 Bücher
über 537 Zeitschriften

aus folgenden Fachgebieten:

Automobil + Motoren
Bauwesen + Immobilien
Business IT + Informatik
Elektrotechnik + Elektronik
Energie + Nachhaltigkeit
Finance + Banking
Management + Führung
Marketing + Vertrieb
Maschinenbau + Werkstoffe
Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Jetzt informieren

Springer Professional "Technik"

Online-Abonnement

Mit Springer Professional "Technik" erhalten Sie Zugriff auf:

über 67.000 Bücher
über 390 Zeitschriften

aus folgenden Fachgebieten:

Automobil + Motoren
Bauwesen + Immobilien
Business IT + Informatik
Elektrotechnik + Elektronik
Energie + Nachhaltigkeit
Maschinenbau + Werkstoffe

Jetzt Wissensvorsprung sichern!

Jetzt informieren

Springer Professional "Wirtschaft"

Online-Abonnement

Mit Springer Professional "Wirtschaft" erhalten Sie Zugriff auf:

über 67.000 Bücher
über 340 Zeitschriften

aus folgenden Fachgebieten:

Bauwesen + Immobilien
Business IT + Informatik
Finance + Banking
Management + Führung
Marketing + Vertrieb
Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Jetzt informieren

Nur mit Berechtigung zugänglich

XCSF’s parameters are set to \(N=4,000, \varepsilon_0=0.002, \beta=0.1, \delta=0.1, \alpha=1, \theta_{\rm GA}=50, \theta_{\rm del}=\theta_{\rm sub}=20, \chi=1\). The mutation probability for each attribute of the rotating hyper-ellipsoidal structures (center, stretch, and angle) are set to μ = 1/n = 0.2. The center is mutated within the receptive field bounds, the stretches are decreased or increased in size maximally doubling or halving their current size; the angles are uniformly changed by maximally 45°. The initial radius of receptive fields is taken uniformly random from [0.00,1]. GA subsumption is applied. Condensation is applied after 80 % of the learning iterations.

Bishop CM (2006) Pattern recognition and machine learning. Springer, New York, Inc, Secaucus, NJMATH

Butz MV (2005) Kernel-based, ellipsoidal conditions in the real-valued XCS classifier system. Genetic and evolutionary computation conference, GECCO 2005, pp 1835–1842

Butz MV, Goldberg DE, Lanzi PL, Sastry K (2007) Problem solution sustenance in XCS: Markov chain analysis of niche support distributions and the impact on computational complexity. Genet Program Evolvable Mach 8:5–37CrossRef

Butz MV, Kovacs T, Lanzi PL, Wilson SW (2004) Toward a theory of generalization and learning in XCS. IEEE Trans Evol Comput 8:28–46CrossRef

Butz MV, Lanzi PL, Wilson SW (2008) Function approximation with XCS: hyperellipsoidal conditions, recursive least squares, and compaction. IEEE Trans Evol Comput 12:355–376CrossRef

Lanzi PL (1999) An analysis of generalization in the XCS classifier system. Evol Comput 7(2):125–149CrossRef

Lanzi PL (1999) An extension to the XCS classifier system for stochastic environments. Genetic and evolutionary computation conference, GECCO 1999, pp 353–360

Orriols-Puig A, Bernadó-Mansilla E (2009) Evolutionary rule-based systems for imbalanced data sets. Soft Compu Fusion Found Methodol Appl 13:213–225. doi:10.1007/s00500-008-0319-7

Orriols-Puig A, Bernadó-Mansilla E, Goldberg DE, Sastry K, Lanzi PL (2009) Facetwise analysis of xcs for problems with class imbalances. IEEE Trans Evol Comput 13(5):1093–1119CrossRef

10.

Rasmussen CE, Williams CKI (2006) Gaussian processes for machine learning. MIT Press, Cambridge, MAMATH

11.

Schaal S, Atkeson CG (1998) Constructive incremental learning from only local information. Neural Comput 10:2047–2084CrossRef

12.

Stalph PO, Butz MV (2011) Learning local linear jacobians for flexible and adaptive robot arm control. Genetic programming and evolvable machines, pp 1–21. doi:10.1007/s10710-011-9147-0

13.

Stalph PO, Butz MV (2010) How fitness estimates interact with reproduction rates: towards variable offspring set sizes in xcsf. In: Bacardit J, Browne W, Drugowitsch J, Bernadó-Mansilla E, Butz MV (eds) Learning classifier systems, volume 6471 of lecture notes in computer science, Springer, Berlin/Heidelberg, pp 47–56

14.

Stalph PO, Llorà X, Goldberg DE, Butz MV (2010) Resource management and scalability of the XCSF learning classifier system. Theor Comput Sci. doi:10.1016/j.tcs.2010.07.007

15.

Stalph PO, Rubinsztajn J, Sigaud O, Butz MV (2010) A comparative study: function approximation with LWPR and XCSF. In: Proceedings of the 12th annual conference on genetic and evolutionary computation, ACM, New York, pp 1863–1870

16.

Vijayakumar S, D’ Souza A, Schaal S (2005) Incremental online learning in high dimensions. Neural Comput 17:2602–2634MathSciNetCrossRef

17.

Wilson SW (1995) Classifier fitness based on accuracy. Evol Comput 3(2):149–175CrossRef

18.

Wilson SW (2002) Classifiers that approximate functions. Nat Comput 1:211–234MathSciNetMATHCrossRef

Titel: XCSF with local deletion: preventing detrimental forgetting
verfasst von: Martin V. Butz
Olivier Sigaud
Publikationsdatum: 01.06.2012
Verlag: Springer-Verlag
Erschienen in: Evolutionary Intelligence / Ausgabe 2/2012
Print ISSN: 1864-5909
Elektronische ISSN: 1864-5917
DOI: https://doi.org/10.1007/s12065-012-0077-4

Springer Professional

Abstract

Bitte loggen Sie sich ein, um Zugang zu Ihrer Lizenz zu erhalten.

Sie haben noch keine Lizenz? Dann Informieren Sie sich jetzt über unsere Produkte:

Springer Professional "Wirtschaft+Technik"

Springer Professional "Technik"

Springer Professional "Wirtschaft"

Weitere Artikel der Ausgabe 2/2012

Function approximation with LWPR and XCSF: a comparative study

On principal component analysis for high-dimensional XCSR

Analysing BioHEL using challenging boolean functions

On XCSR for electronic fraud detection

Production system rules as protein complexes from genetic regulatory networks: an initial study

Special issue on advances in learning classifier systems

Premium Partner