Interpretability and complexity of design in the creation of fuzzy logic systems — a user study

Razak, T.R., Garibaldi, J.M., Wagner, C., Pourabdollah, A. ORCID: 0000-0001-7737-1393 and Soria, D., 2018. Interpretability and complexity of design in the creation of fuzzy logic systems — a user study. In: S. Sundaram, ed., Proceedings of the 2018 IEEE Symposium Series on Computational Intelligence (SSCI 2018), Bangalore, India, 18-21 November 2018. Piscataway, NJ: Institute of Electrical and Electronics Engineers, pp. 420-426. ISBN 9781538692769

Preview

Text 13810_Pourabdollah.pdf - Post-print Download (5MB) | Preview

Abstract

In recent years, researchers have become increasingly more interested in designing an interpretable Fuzzy Logic System (FLS). Many studies have claimed that reducing the complexity of FLSs can lead to improved model interpretability. That is, reducing the number of rules tends to reduce the complexity of FLSs, thus improving their interpretability. However, none of these studies have considered interpretability and complexity from human perspectives. Since interpretability is of a subjective nature, it is essential to see how people perceive interpretability and complexity particularly in relation to creating FLSs. Therefore, in this paper we have investigated this issue using an initial user study. This is the first time that a user study has been used to assess the interpretability and complexity of designs in relation to creating FLSs. The user study involved a range of expert practitioners in FLSs and received a diverse set of answers. We are interested to see whether, from the perspectives of people, FLSs are necessarily more interpretable when they are less complex in terms of their design. Although the initial user study is based on small samples (i.e., 25 participants), nevertheless this research provides initial insight into this issue that motivates our future research.

Item Type:	Chapter in book
Creators:	Razak, T.R., Garibaldi, J.M., Wagner, C., Pourabdollah, A. and Soria, D.
Publisher:	Institute of Electrical and Electronics Engineers
Place of Publication:	Piscataway, NJ
Date:	November 2018
ISBN:	9781538692769
Identifiers:	Number Type 10.1109/ssci.2018.8628924 DOI
Divisions:	Schools > School of Science and Technology
Record created by:	Jonathan Gallacher
Date Added:	01 May 2019 10:45
Last Modified:	01 May 2019 10:45
URI:	https://irep.ntu.ac.uk/id/eprint/36369

Actions (login required)

Edit View

Views

Downloads