A stochastic model for analyzing the interpretability-accuracy trade-off in interpretable fuzzy systems using nested Hyperball structures

Krisztián Balázs, L. Kóczy

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Our recent work proposed a new meaning preservation approach together with a parameterizable nested hyperball structured search space for interpretable fuzzy systems in order to solve a problem of inconsistency observed in conventional interpretable fuzzy knowledge bases and simultaneously to address the adjustment of the trade-off between interpretability and accuracy. Based on intuitive reasonings and simulation results a conjecture was formulated about favorable trade-off adjustment properties of the proposed method. The aim of the present paper is to construct a mathematical model, in which the conjectured properties can be analyzed and formally verified. Some computational considerations about the interpretation of the resulting knowledge bases are also made.

Original languageEnglish
Title of host publication8th Conference of the European Society for Fuzzy Logic and Technology, EUSFLAT 2013 - Advances in Intelligent Systems Research
Pages72-79
Number of pages8
Volume32
Publication statusPublished - 2013
Event8th Conference of the European Society for Fuzzy Logic and Technology, EUSFLAT 2013 - Milan, Italy
Duration: Sep 11 2013Sep 13 2013

Other

Other8th Conference of the European Society for Fuzzy Logic and Technology, EUSFLAT 2013
CountryItaly
CityMilan
Period9/11/139/13/13

Fingerprint

Fuzzy systems
Stochastic models
Mathematical models

Keywords

  • Formal analysis
  • Interpretability-accuracy trade-off
  • Interpretable fuzzy systems
  • Knowledge extraction

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Information Systems

Cite this

Balázs, K., & Kóczy, L. (2013). A stochastic model for analyzing the interpretability-accuracy trade-off in interpretable fuzzy systems using nested Hyperball structures. In 8th Conference of the European Society for Fuzzy Logic and Technology, EUSFLAT 2013 - Advances in Intelligent Systems Research (Vol. 32, pp. 72-79)

A stochastic model for analyzing the interpretability-accuracy trade-off in interpretable fuzzy systems using nested Hyperball structures. / Balázs, Krisztián; Kóczy, L.

8th Conference of the European Society for Fuzzy Logic and Technology, EUSFLAT 2013 - Advances in Intelligent Systems Research. Vol. 32 2013. p. 72-79.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Balázs, K & Kóczy, L 2013, A stochastic model for analyzing the interpretability-accuracy trade-off in interpretable fuzzy systems using nested Hyperball structures. in 8th Conference of the European Society for Fuzzy Logic and Technology, EUSFLAT 2013 - Advances in Intelligent Systems Research. vol. 32, pp. 72-79, 8th Conference of the European Society for Fuzzy Logic and Technology, EUSFLAT 2013, Milan, Italy, 9/11/13.
Balázs K, Kóczy L. A stochastic model for analyzing the interpretability-accuracy trade-off in interpretable fuzzy systems using nested Hyperball structures. In 8th Conference of the European Society for Fuzzy Logic and Technology, EUSFLAT 2013 - Advances in Intelligent Systems Research. Vol. 32. 2013. p. 72-79
Balázs, Krisztián ; Kóczy, L. / A stochastic model for analyzing the interpretability-accuracy trade-off in interpretable fuzzy systems using nested Hyperball structures. 8th Conference of the European Society for Fuzzy Logic and Technology, EUSFLAT 2013 - Advances in Intelligent Systems Research. Vol. 32 2013. pp. 72-79
@inproceedings{3879b3fc319048738eafff7ef0acb498,
title = "A stochastic model for analyzing the interpretability-accuracy trade-off in interpretable fuzzy systems using nested Hyperball structures",
abstract = "Our recent work proposed a new meaning preservation approach together with a parameterizable nested hyperball structured search space for interpretable fuzzy systems in order to solve a problem of inconsistency observed in conventional interpretable fuzzy knowledge bases and simultaneously to address the adjustment of the trade-off between interpretability and accuracy. Based on intuitive reasonings and simulation results a conjecture was formulated about favorable trade-off adjustment properties of the proposed method. The aim of the present paper is to construct a mathematical model, in which the conjectured properties can be analyzed and formally verified. Some computational considerations about the interpretation of the resulting knowledge bases are also made.",
keywords = "Formal analysis, Interpretability-accuracy trade-off, Interpretable fuzzy systems, Knowledge extraction",
author = "Kriszti{\'a}n Bal{\'a}zs and L. K{\'o}czy",
year = "2013",
language = "English",
isbn = "9781629932194",
volume = "32",
pages = "72--79",
booktitle = "8th Conference of the European Society for Fuzzy Logic and Technology, EUSFLAT 2013 - Advances in Intelligent Systems Research",

}

TY - GEN

T1 - A stochastic model for analyzing the interpretability-accuracy trade-off in interpretable fuzzy systems using nested Hyperball structures

AU - Balázs, Krisztián

AU - Kóczy, L.

PY - 2013

Y1 - 2013

N2 - Our recent work proposed a new meaning preservation approach together with a parameterizable nested hyperball structured search space for interpretable fuzzy systems in order to solve a problem of inconsistency observed in conventional interpretable fuzzy knowledge bases and simultaneously to address the adjustment of the trade-off between interpretability and accuracy. Based on intuitive reasonings and simulation results a conjecture was formulated about favorable trade-off adjustment properties of the proposed method. The aim of the present paper is to construct a mathematical model, in which the conjectured properties can be analyzed and formally verified. Some computational considerations about the interpretation of the resulting knowledge bases are also made.

AB - Our recent work proposed a new meaning preservation approach together with a parameterizable nested hyperball structured search space for interpretable fuzzy systems in order to solve a problem of inconsistency observed in conventional interpretable fuzzy knowledge bases and simultaneously to address the adjustment of the trade-off between interpretability and accuracy. Based on intuitive reasonings and simulation results a conjecture was formulated about favorable trade-off adjustment properties of the proposed method. The aim of the present paper is to construct a mathematical model, in which the conjectured properties can be analyzed and formally verified. Some computational considerations about the interpretation of the resulting knowledge bases are also made.

KW - Formal analysis

KW - Interpretability-accuracy trade-off

KW - Interpretable fuzzy systems

KW - Knowledge extraction

UR - http://www.scopus.com/inward/record.url?scp=84891750758&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84891750758&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84891750758

SN - 9781629932194

VL - 32

SP - 72

EP - 79

BT - 8th Conference of the European Society for Fuzzy Logic and Technology, EUSFLAT 2013 - Advances in Intelligent Systems Research

ER -