Research Repository

Enhanced Type-2 Wang-Mendel Approach

Gupta, Prashant K and Andreu-Perez, Javier (2022) 'Enhanced Type-2 Wang-Mendel Approach.' Journal of Experimental and Theoretical Artificial Intelligence. pp. 1-26. ISSN 0952-813X (In Press)

[img] Text
WMIT2_JETAI.pdf - Accepted Version
Restricted to Repository staff only until 1 January 2100.
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (779kB) | Request a copy

Abstract

The Wang-Mendel Approach (WMA) focuses on combining the numerical as well as linguistic information for achieving greater explainability for inference models. The standard WMA models the linguistic information using type-1 (T1) fuzzy sets (FSs), which have a reduced capability to model the semantics of linguistic information. Therefore, we propose a novel Enhanced WMA, which models the linguistic information using the type-2 (T2) FSs. Further, our Enhanced T2 FS based WMA can be modified to reflect the use of interval type-2 (IT2) FSs, for modeling linguistic uncertainty. IT2 FSs are suitable when better uncertainty handling capabilities are required compared to T1 FSs, however, at a computational cost lesser than the T2 FSs. Performance of Enhanced WMA is demonstrated through a real-world crop-yield prediction problem in smart agriculture and an additional exemplar application on users' satisfaction ratings. Further, we have compared our approach with the performance obtained from the T1 FS based WMA and the original estimations given in the original data. We found that our Enhanced WMA achieves better precision than the other two with 95% confidence level. To the best of our knowledge, no one has proposed the use of T2 FSs for modeling linguistic uncertainty in the WMA before.

Item Type: Article
Divisions: Faculty of Science and Health
Faculty of Science and Health > Computer Science and Electronic Engineering, School of
SWORD Depositor: Elements
Depositing User: Elements
Date Deposited: 28 Sep 2022 14:17
Last Modified: 20 Nov 2022 09:48
URI: http://repository.essex.ac.uk/id/eprint/33318

Actions (login required)

View Item View Item