arxiv: 2605.00048 · v1 · submitted 2026-04-29 · 🧮 math.GM

Recognition: unknown

Hierarchical similarity-based approximate reasoning with restricted equivalence function

Dechao Li, Yuhui Zhu

Pith reviewed 2026-05-09 20:59 UTC · model grok-4.3

classification 🧮 math.GM

keywords restricted equivalence functionssimilarity-based approximate reasoninghierarchical fuzzy reasoningfuzzy rule explosionaggregation functionsapproximation equalityRaha's SBAR

0 comments

The pith

Integrating restricted equivalence functions into hierarchical similarity-based approximate reasoning limits fuzzy rule explosion while preserving approximation quality.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper seeks to extend Raha's similarity-based approximate reasoning by incorporating restricted equivalence functions as similarity measures within a hierarchical structure. It first characterizes REFs in terms of aggregation functions, then verifies that the hierarchical versions maintain the approximation equality of the original method. The central proposal is two specific REF-based hierarchical SBAR approaches designed to address the rapid growth of fuzzy rules as system complexity increases.

Core claim

The paper constructs two REF-based hierarchical versions of Raha's SBAR method. After characterizing REFs via a given aggregation function and confirming that approximation equality holds under this integration, the work shows that the hierarchical organization restrains the explosion of fuzzy rules without altering the core inference behavior of the original SBAR system.

What carries the argument

Restricted equivalence functions (REFs) serving as similarity measures inside a hierarchical extension of Raha's SBAR method, which organizes the rule base into layers to control size.

If this is right

The two hierarchical methods preserve the approximation equality property of the original Raha SBAR.
REFs characterized through aggregation functions enable consistent similarity evaluation across the layers.
The hierarchical organization directly reduces the total number of fuzzy rules needed for a given reasoning task.
The construction applies to any REF that satisfies the stated characterization with an aggregation function.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same layering technique could be tested on other similarity measures beyond REFs to see whether rule reduction generalizes.
In applied fuzzy control or decision systems, the reduced rule count might lower memory and computation costs enough to handle larger input spaces.
One could compare the inference speed of these hierarchical methods against flat SBAR on benchmark problems with increasing numbers of variables.

Load-bearing premise

That REFs can measure similarity between fuzzy sets in a manner that keeps the approximation equality of the original SBAR intact once the structure becomes hierarchical.

What would settle it

An explicit pair of input fuzzy sets for which the output of one of the proposed hierarchical REF-based SBAR methods deviates from the output produced by the non-hierarchical Raha SBAR on the same inputs.

read the original abstract

Given that the restricted equivalence functions (REFs) can serve to measure the similarity of two fuzzy sets, this motivates the integration of REFs with similarity-based approximate reasoning systems to enhance inference capabilities. Therefore, this work primarily constructs hierarchical similarity-based approximate reasoning (SBAR) using REFs. Specifically, we first characterize REFs with a given aggregation function, then discuss the approximation equality of SBAR method proposed by Raha et al. with REFs. Finally, we suggest two REF-based hierarchical Raha's SBAR methods which efficiently restrain the explosion of fuzzy rules.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Li and Zhu give two hierarchical REF-based extensions to Raha's SBAR to cut rule explosion, but the key approximation equality needs explicit checks in the hierarchy.

read the letter

The main thing to know is that Li and Zhu have developed two hierarchical versions of Raha's similarity-based approximate reasoning by incorporating restricted equivalence functions, with the aim of restraining the growth of fuzzy rules while keeping the approximation intact. They characterize REFs using a given aggregation function, discuss the approximation equality from the original SBAR when using REFs, and then propose the two new hierarchical methods. This is new in applying hierarchy to the REF-integrated SBAR. It does well by tackling the practical problem of rule explosion in fuzzy systems through a structured approach that organizes the reasoning. The work is constructive and stays grounded in the existing literature on SBAR. It provides a clear path for extending the method. The soft spot lies in confirming that the approximation equality holds after the hierarchical modification. The aggregation used to combine similarities at different levels must satisfy properties that allow the equality to carry over from the flat case. The paper discusses this, but if it lacks detailed proofs or examples demonstrating preservation, then the claim relies partly on unverified assumptions about the REF composition. This is a moderate issue that could be addressed with more analysis. This paper is for specialists in fuzzy logic and approximate reasoning who work with large rule bases. A reader familiar with Raha's method will see value in the specific hierarchical constructions and the REF integration. It deserves serious peer review because it offers a targeted improvement in an established area, even though the novelty is incremental and some details on the equality may need strengthening.

Referee Report

2 major / 2 minor

Summary. The manuscript integrates restricted equivalence functions (REFs) as a similarity measure into similarity-based approximate reasoning (SBAR). It first characterizes REFs via a given aggregation function, discusses the approximation equality of Raha et al.'s original SBAR method when REFs are substituted for the similarity measure, and then proposes two hierarchical REF-based extensions of Raha's SBAR, claiming these constructions efficiently restrain the explosion of fuzzy rules while preserving the original approximation properties.

Significance. If the hierarchical constructions are shown to preserve the approximation equality of the flat SBAR method while demonstrably reducing rule count, the work would offer a practical extension for scalable fuzzy inference. The constructive approach that builds directly on cited prior SBAR literature is a positive feature; however, the absence of explicit derivations, proofs, or counter-example checks for the hierarchical case limits the result's immediate utility.

major comments (2)

[Section discussing approximation equality of SBAR with REFs] The central claim that the two REF-based hierarchical SBAR methods preserve the approximation equality of Raha et al.'s original method requires a demonstration that recursive or tree-structured application of the REF (combined via the aggregation function) commutes with the approximation operator in the same manner as the single-level case. The manuscript states that the equality is 'discussed' but supplies no general proof or counter-example verification for the hierarchical setting; without this, the claim that the equality carries over is unsupported.
[Section presenting the two REF-based hierarchical Raha's SBAR methods] The assertion that the hierarchical methods 'efficiently restrain the explosion of fuzzy rules' is load-bearing for the contribution. The manuscript must supply either a quantitative comparison of rule cardinality (flat vs. hierarchical) or a formal bound on rule growth under the REF-based similarity; absent such analysis, the efficiency claim remains qualitative.

minor comments (2)

[Abstract] The abstract would be strengthened by a single sentence or equation indicating how the aggregation function is used to combine level-wise REF similarities.
[Section on hierarchical constructions] Notation for the hierarchical structure (e.g., how the tree levels are indexed and how the final approximation is obtained) should be introduced explicitly before the constructions are presented.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which highlight areas where the manuscript can be strengthened. We address each major comment below and will revise the manuscript to incorporate explicit proofs and quantitative analysis.

read point-by-point responses

Referee: [Section discussing approximation equality of SBAR with REFs] The central claim that the two REF-based hierarchical SBAR methods preserve the approximation equality of Raha et al.'s original method requires a demonstration that recursive or tree-structured application of the REF (combined via the aggregation function) commutes with the approximation operator in the same manner as the single-level case. The manuscript states that the equality is 'discussed' but supplies no general proof or counter-example verification for the hierarchical setting; without this, the claim that the equality carries over is unsupported.

Authors: We agree that an explicit demonstration is required for the hierarchical case. The manuscript discusses the single-level approximation equality in detail and constructs the hierarchical methods recursively from the same REF and aggregation functions. In the revised manuscript, we will add a theorem proving preservation by induction on hierarchy depth, showing that the recursive application commutes with the approximation operator. We will also include a small-scale counter-example verification to illustrate the result. revision: yes
Referee: [Section presenting the two REF-based hierarchical Raha's SBAR methods] The assertion that the hierarchical methods 'efficiently restrain the explosion of fuzzy rules' is load-bearing for the contribution. The manuscript must supply either a quantitative comparison of rule cardinality (flat vs. hierarchical) or a formal bound on rule growth under the REF-based similarity; absent such analysis, the efficiency claim remains qualitative.

Authors: We acknowledge that the efficiency claim is currently stated qualitatively. In the revision, we will add a dedicated analysis section providing a formal bound on rule cardinality for the hierarchical structures (showing polynomial reduction relative to the flat case) along with a concrete numerical comparison for a sample fuzzy rule base. revision: yes

Circularity Check

0 steps flagged

No significant circularity; constructive extension of prior SBAR

full rationale

The paper first characterizes REFs via an aggregation function (external to the target result), discusses the approximation equality from Raha et al. as a cited prior result, and then proposes two new hierarchical constructions. No equation or claim reduces a 'prediction' or equality to a quantity defined inside the paper by construction, nor does any load-bearing step rely on a self-citation chain that itself lacks independent verification. The work is self-contained against the external benchmark of Raha's flat SBAR and standard REF properties.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available; no explicit free parameters, axioms, or invented entities can be extracted. The central claim rests on the unstated assumption that REFs are suitable similarity measures for fuzzy sets and that hierarchical composition preserves the properties of the base SBAR method.

pith-pipeline@v0.9.0 · 5381 in / 1077 out tokens · 31878 ms · 2026-05-09T20:59:24.922962+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

45 extracted references

[1]

Aguil´ o, J

I. Aguil´ o, J. Su¯ ner, J. Torrens, New types of contrapositivis ation of fuzzy implications with respect to fuzzy negations, Information Sciences 322(2015 )223-236

2015
[2]

Altalhi, J.I

A.H. Altalhi, J.I. Forc´ en, M. Pagola et al., Moderate deviation and r estricted equivalence functions for measuring similarity between data, Information Scien ces 501(2019)19-29

2019
[3]

Baczy´ nski, B

M. Baczy´ nski, B. Jayaram, Fuzzy Implications, Springer, Berlin , 2008

2008
[4]

I. Beg, S. Ashraf, Fuzzy similarity and measure of similarity with /suppress Lukasiewicz implicator, New Mathematics and Natural Computation 4(2)(2008)191-206

2008
[5]

Bustince, E

H. Bustince, E. Barrenechea, M. Pagola, Restricted equivalenc e functions, Fuzzy Sets and Systems 157(2006)2333-2346

2006
[6]

Bustince, E

H. Bustince, E. Barrenechea, M. Pagola, Image thresholding us ing restricted equiv- alence functions and maximizing the measures of similarity, Fuzzy Set s and Systems 158(5)(2007)496-516

2007
[7]

Bustince, E

H. Bustince, E. Barrenechea, M. Pagola, Relationship between r estricted dissimilarity functions, restricted equivalence functions and normal EN -functions: image thresholding invariant, Pattern Recognition Letters 29(2008)525-536

2008
[8]

Bustince, M.J

H. Bustince, M.J. Campi´ on, L. De Miguel, E. Indur´ ain, Strong ne gations and restricted equivalence functions revisited: Ananalytical and topological appr oach, Fuzzy Sets and Systems 441(2022)110-129

2022
[9]

Bustince, C

H. Bustince, C. Marco-Detchart, J. Fernandez et al., Similarity b etween interval-valued fuzzy sets taking into account the width of the intervals and admiss ible orders, Fuzzy Sets and Systems 390(2020)23-47

2020
[10]

Cornelis, M

C. Cornelis, M. D. Cock, E. Kerre, Eﬃcient approximate reason ing with positive and negative information, Heidelberg, Germany: Springer-Verlag, 200 4, vol. 3214, KES 2004, LNAI, pp. 779-785

2004
[11]

Chaira, Intuitionistic fuzzy segmentation of medical images, IEEE Transactions on Biomedical Engineering 57(6)(2010)1430-1436

T. Chaira, Intuitionistic fuzzy segmentation of medical images, IEEE Transactions on Biomedical Engineering 57(6)(2010)1430-1436. 26

2010
[12]

Chen, M.S Yeh, P.Y

S.M. Chen, M.S Yeh, P.Y. Hsiao, A comparison of similarity measures of fuzzy values, Fuzzy Sets and Systems 72(1995)79-89

1995
[13]

Dai, D.W

S.S. Dai, D.W. Pei, D.H. Guo, Robustness analysis of full implication in ference method, International Journal of Approximate Reasoning 54(2013)653- 666

2013
[14]

De Miguel, R

L. De Miguel, R. Santiago, C. Wagner et al., Extension of restrict ed equivalence func- tions and similarity measures for type-2 fuzzy sets, IEEE Transac tions on Fuzzy Systems 30(9)(2022)4005-4016

2022
[15]

De Baets, H

B. De Baets, H. De Meyer, Transitivity-preserving fuzziﬁcatio n schemes for cardinality- based similarity measures, European Journal of Operational Rese arch 160 (2005)726-740

2005
[16]

Demirli, I.B

K. Demirli, I.B. Turksen, Rule break up with compositional rule of in ference, in Proc. FUZZ-IEEE’92, San Diego, 1992, pp. 949-956

1992
[17]

Deng, Y.L

G.N. Deng, Y.L. Jiang, Fuzzy reasoning method by optimizing the s imilarity of truth- tables, Information Sciences 288(2014)290-313

2014
[18]

Dimuro, B

G.P. Dimuro, B. Bedregal, J. Fernandez et al., The law of O-conditionality for fuzzy implications constructed from overlap and grouping functions, Int ernational Journal of Approximate Reasoning 105(2019)27-48

2019
[19]

Fan, W.X

J.L. Fan, W.X. Xie, Some notes on similarity measure and proximity m easure, Fuzzy Sets and Systems 101(1999)403-412

1999
[20]

Ferrero-Jaurrieta, Z

M. Ferrero-Jaurrieta, Z. Tak´ a˘ c, I. Rodriguez-Martinez et al., From restricted equivalence functions on Ln to similarity measures between fuzzy multisets, IEEE Transactions on Fuzzy Systems 31(8)(2023)2709-2721

2023
[21]

Fodor, M

J. Fodor, M. Roubens, Fuzzy Preference Modelling and Multicrit eria Decision Support, in: Theory and Decision Library, Kluwer Academic Publishers, Dordre cht, 1994

1994
[22]

Grabisch, J.L

M. Grabisch, J.L. Marichal, R. Mesiar, E. Pap, Aggregation Func tions, Cambridge Uni- versity Press, New York, 2009

2009
[23]

Jayaram, On the law of importation ( x∧ y)→ z = x→ (y→ z) in fuzzy logic, IEEE Transactions on Fuzzy Systems 16(2008)130-144

B. Jayaram, On the law of importation ( x∧ y)→ z = x→ (y→ z) in fuzzy logic, IEEE Transactions on Fuzzy Systems 16(2008)130-144

2008
[24]

Kaleva, S

O. Kaleva, S. Seikkala, On fuzzy metric spaces, Fuzzy Sets and Systems 12(3)(1984)215- 229

1984
[25]

Klement, R

E.P. Klement, R. Mesiar, E. Pap, Triangular Norms, Kluwer Acade mic Publishers, Dor- drecht, Boston, 2000

2000
[26]

D.C. Li, Q.N. Guo, MISO hierarchical inference engine satisfying t he law of importation with aggregation functions, Artiﬁcial Intelligence Review 56(10)(2 023)1-26
[27]

D.C. Li, Z.S. Liu, Q.N. Guo, Hierarchical fuzzy inference based on Bandler-Kohout sub- product, Information Sciences 677(2024)120889. 27

2024
[28]

Y.F. Li, K.Y. Qin, X.X. He, Some new approaches to constructing s imilarity measures, Fuzzy Sets and Systems 234(2014)46-60

2014
[29]

Y.F. Li, K.Y. Qin, X.X. He et al., Properties of Raha’s similarity-based approximate reasoning method, Fuzzy Sets and Systems 294(2016)48-62

2016
[30]

Liu, Entropy, distance measure and similarity measure of fu zzy sets and their rela- tions, Fuzzy Sets and Systems 52(1992)305-318

X.C. Liu, Entropy, distance measure and similarity measure of fu zzy sets and their rela- tions, Fuzzy Sets and Systems 52(1992)305-318

1992
[31]

Lowen, On fuzzy complements, Information Sciences 14(2) (1978)107-113

R. Lowen, On fuzzy complements, Information Sciences 14(2) (1978)107-113

1978
[32]

Luo, Y.J

M.X. Luo, Y.J. Wang, R.R. Zhao, Interval-valued fuzzy reasonin g method based on similarity measure, Journal of Logical and Algebraic Methods in Pr ogramming 113(2020)100541

2020
[33]

Marco-Detchart, J

C. Marco-Detchart, J. Cerron, L. De Miguel et al., A framewor k for radial data comparison and its application to ﬁngerprint analysis, Applied Soft Computing 46( 2016)246-259

2016
[34]

Palmeira, B

E. Palmeira, B. Bedregal, H. Bustince, A generalization of a char acterization theorem of restricted equivalence functions, in: H. Bustince, J. Fernandez, R. Mesiar, T. Calvo (Eds.), Aggregation Functions in Theory and in Practise, Vol. 228 of Advanc es in Intelligent Systems and Computing, Springer Berlin Heidelberg, 2013, pp. 453- 464

2013
[35]

Palmeira, B

E. Palmeira, B. Bedregal, H. Bustince et al., Application of two diﬀe rent methods for extending lattice-valued restricted equivalence functions used fo r constructing similarity measures on L-fuzzy sets, Information Sciences 441(2018)95- 112

2018
[36]

Pappis and N

C. Pappis and N. Karacapilidis, A comparative assessment of mea sures of similarity of fuzzy values, Fuzzy Sets and Systems 56(1993)171-174

1993
[37]

Qiao, Restricted equivalence functions induced from fuzzy im plication functions, Ira- nian Journal of Fuzzy Systems 20(2)(2023)151-160

J. Qiao, Restricted equivalence functions induced from fuzzy im plication functions, Ira- nian Journal of Fuzzy Systems 20(2)(2023)151-160

2023
[38]

Qin, J.L

K.Y. Qin, J.L. Yang, Z.C. Liu, On the similarity property of some fuz zy reasoning meth- ods, Journal of Intelligent and Fuzzy Systems 33(2017)2291-23 03

2017
[39]

S. Raha, A. Hossain, S. Ghosh, Similarity based approximate rea soning: fuzzy control, Journal of Applied Logic 6(2008)47-71

2008
[40]

Turksen, Z

I.B. Turksen, Z. Zhong, An approximate analogical reasoning a pproach based on similarity measures, IEEE Trans. Syst. Man Cybern. 18(1988)1049-1056

1988
[41]

Wang, A Course in Fuzzy Systems and Control, Prentice Hall PTR, Upper Saddle River, 1997

L.X. Wang, A Course in Fuzzy Systems and Control, Prentice Hall PTR, Upper Saddle River, 1997

1997
[42]

Wang, B.D

X.Z. Wang, B.D. Baets, E. Kerre, A comparative study of similarit y measures, Fuzzy Sets and Systems 73(1995)259-268. 28

1995
[43]

Wang, J.Y

G.J. Wang, J.Y. Duan, On robustness of the full implication triple I inference method with respect to ﬁner measurements, International Journal of A pproximate Reasoning 55(2014)787-796

2014
[44]

Wang, Y.P

D.G. Wang, Y.P. Meng, H.X. Li, A fuzzy similarity inference method f or fuzzy reasoning, Computers and Mathematics with Applications 56(2008)2445-2454

2008
[45]

Yearsley, A

J.M. Yearsley, A. Barque-Duran, E. Scerrati et al., The triang le inequality constraint in similarity judgments, Progress in Biophysics and Molecular Biology 130 (2017)26-32. 29

2017