arxiv: 2605.00045 · v1 · submitted 2026-04-29 · 🧮 math.GM

Recognition: unknown

Measuring and aggregating {ε}-T-transitive fuzzy relations

Dechao Li , Yutao Yao , Jingyao Duan

Authors on Pith no claims yet

Pith reviewed 2026-05-09 21:10 UTC · model grok-4.3

classification 🧮 math.GM

keywords fuzzy relationsT-transitivityε-T-transitivityaggregation functionsfuzzy implicationsclusteringfuzzy inference

0 comments

The pith

ε-T-transitive fuzzy relations allow clustering and inference with a small permissible error instead of requiring full T-transitivity.

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

This paper investigates two measures of the degree to which a fuzzy relation satisfies T-transitivity, each built from standard fuzzy implications. It defines ε-T-transitive fuzzy relations as those whose measured transitivity falls short by at most a chosen tolerance ε. The work then identifies which aggregation functions keep the resulting relation ε-T-transitive and shows how such relations support direct inference steps and object clustering. A sympathetic reader would care because many real-world fuzzy relations never reach exact T-transitivity, so a controlled-error substitute avoids the artificial strengthening imposed by transitive-closure methods.

Core claim

The paper introduces the concept of an ε-T-transitive fuzzy relation, characterizes the aggregation functions that preserve the ε-T-transitivity of fuzzy relations, and utilizes the ε-T-transitive fuzzy relation to make inferences and cluster objects, arguing that this approach is reasonable under a permissible error ε compared with computing the T-transitive closure.

What carries the argument

The ε-T-transitive fuzzy relation, constructed from two T-transitivity measures based on fuzzy implications, which quantifies closeness to full transitivity and permits a tunable deviation ε while still supporting aggregation, inference, and clustering.

If this is right

Aggregation functions that preserve ε-T-transitivity can combine several such relations into one without losing the controlled-error property.
Inferences about object relations can be performed directly on ε-T-transitive fuzzy relations.
Clustering of objects can be carried out using ε-T-transitive relations, offering a practical alternative to first computing the full T-transitive closure.

Where Pith is reading between the lines

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

The same tolerance idea could be applied to other approximate properties of fuzzy relations, such as approximate reflexivity or symmetry.
In large-scale decision systems, tuning ε according to observed data noise might reduce the computational burden of closure calculations while maintaining usable accuracy.
The framework suggests testing whether ε-T-transitivity yields more stable clusters than strict closure when input relations contain measurement errors.

Load-bearing premise

That the two chosen measures based on well-known fuzzy implications correctly capture the intuitive degree of T-transitivity, and that allowing a small permissible error ε is acceptable and useful in the target clustering and inference applications.

What would settle it

A side-by-side clustering experiment on a dataset with known ground-truth groups, comparing partitions obtained from ε-T-transitive relations at varying ε against partitions from the exact T-transitive closure, to check whether the approximate versions recover the same groups.

read the original abstract

The transitivity of fuzzy relations plays an important role in fuzzy set theory, artificial intelligence, clustering and decision-making. However, it is often difficult for fuzzy relations to satisfy the transitivity property in many practical applications. This has motivated researchers to investigate the degree to which a fuzzy relation is transitive. Therefore, this work first investigates two different measures of T-transitivity for fuzzy relations using some well-known fuzzy implications. And then, the relationship between two different degrees of transitivity is investigated. Further, the concept of an {\epsilon}-T-transitive fuzzy relation is introduced, and the aggregation functions that preserve the {\epsilon}-T-transitivity of fuzzy relations are characterized. Finally, the {\epsilon}-T-transitive fuzzy relation is utilized to make inferences and cluster objects. Compared to finding the T-transitive closure, it is reasonable to cluster objects using the {\epsilon}-T-transitive fuzzy relation under the permissible error.

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

This paper defines an ε-relaxation of T-transitivity for fuzzy relations, gives two measures based on standard implications, characterizes the aggregations that preserve the relaxed property, and positions it as a simpler alternative to transitive closure for clustering.

read the letter

The core contribution is the introduction of ε-T-transitive fuzzy relations together with the explicit characterization of which aggregation functions preserve that property. They start from two transitivity-degree measures built on common fuzzy implications, establish the relationship between those measures, then move to the ε version and its preservation results before sketching uses in inference and clustering. The characterization step is the part that actually adds something not already in the cited T-transitivity literature, and it stays within the usual technical patterns of the field rather than requiring new machinery. The suggestion that allowing a controlled ε makes clustering easier than computing the full closure is presented as a modeling choice rather than a theorem, which keeps the claim modest and defensible on its own terms. The abstract lays out a clean sequence from definitions to applications, and the stress-test note confirms there are no hidden assumptions about continuity or specific t-norms that would break the internal logic. The weakest part is the practical claim: the two chosen measures are standard, so their ability to capture an intuitive degree of transitivity rests on how well those implications align with the target applications, and the abstract gives no concrete examples or side-by-side comparisons to show the ε approach is meaningfully simpler or more accurate in practice. That gap is minor for a theoretical note but would matter if the paper aims at applied readers. This is for fuzzy-set theorists and people who already work with relations in clustering or decision procedures under uncertainty. A reader who needs a clean way to handle approximate transitivity in aggregation steps will find the preservation results useful; someone looking for broad methodological shifts or heavy empirical validation will not. The work is coherent enough on its own terms to deserve referee time rather than a desk reject, even if the applications section may need more detail after review.

Referee Report

0 major / 3 minor

Summary. The manuscript defines two measures of T-transitivity degree for fuzzy relations via standard fuzzy implications, establishes the relationship between these measures, introduces the relaxed notion of ε-T-transitive fuzzy relations, characterizes the aggregation functions that preserve ε-T-transitivity, and applies the concept to inference and clustering, positioning it as a practical alternative to computing the full T-transitive closure when a permissible error ε is acceptable.

Significance. If the characterizations and preservation results hold, the work supplies a flexible, application-oriented relaxation of transitivity that avoids the computational overhead of transitive closure while retaining enough structure for clustering and inference tasks. The aggregation-function characterization is a concrete, reusable contribution within fuzzy-relation theory and could support systematic construction of approximate relations without extra continuity or boundedness assumptions.

minor comments (3)

[Abstract] The abstract states that the measures are based on 'some well-known fuzzy implications' but does not name them; listing the specific implications (e.g., Łukasiewicz, Gödel) in the abstract would improve immediate readability.
[Definitions and measures section] Notation for the two transitivity-degree measures should be introduced with an explicit equation number at first use and then referenced consistently in the characterization of preserving aggregators.
[Applications section] The clustering and inference examples would benefit from a small table or numerical illustration showing how ε-T-transitivity differs from the strict transitive closure on a concrete relation.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the careful reading and the positive evaluation of our manuscript. The referee's summary and significance assessment accurately reflect the paper's contributions on measures of T-transitivity, the introduction of ε-T-transitive fuzzy relations, and the characterization of aggregation functions that preserve this property. We are pleased with the recommendation for minor revision.

Circularity Check

0 steps flagged

No significant circularity; new definitions and characterizations are independent

full rationale

The paper introduces two measures of T-transitivity for fuzzy relations based on standard well-known fuzzy implications, proves a relationship between those measures, defines the new concept of ε-T-transitive fuzzy relations, characterizes the aggregation functions that preserve ε-T-transitivity, and applies the notion to inference and clustering. All steps rely on internal proofs within fuzzy-relation theory and do not reduce any claimed result to a fitted parameter, self-defined quantity, or self-citation chain. The ε-relaxation is presented explicitly as a modeling choice for practical applications rather than a derived theorem forced by prior content. No load-bearing step collapses to an input by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 1 invented entities

The work rests on the standard algebraic properties of triangular norms and fuzzy implications taken from the existing literature, plus the new definitional extension to ε-T-transitivity; no data-fitted parameters are introduced.

axioms (2)

standard math Standard properties of triangular norms (T-norms) and fuzzy implications hold
Invoked when defining the two measures of T-transitivity
domain assumption A fuzzy relation is a mapping from X × X to the unit interval [0,1]
Basic setup for all subsequent definitions

invented entities (1)

ε-T-transitive fuzzy relation no independent evidence
purpose: Relaxed transitivity allowing a controlled deviation ε
New definition introduced to enable practical clustering and inference

pith-pipeline@v0.9.0 · 5457 in / 1455 out tokens · 44251 ms · 2026-05-09T21:10:42.450429+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

46 extracted references

[1]

Alavi, N

S.M. Alavi, N. Khazravi, Evaluation and ranking of fuzzy sets under equivalence fuzzy relations asα-certainty andβ-possibility, Expert Systems with Applications 248 (2024)123175

2024
[2]

Alc´ antara, T

M.S.S. Alc´ antara, T. Dias, W.R. Oliveira et al., A survey of categorical properties of L-fuzzy relations, Fuzzy Sets and Systems 425(2021)62-82

2021
[3]

Amini, N

A. Amini, N. Firouzkouhi, M. Nazari et al., Artificial intelligent Global Online Learn- ing (GOL) theory by generalizedn-ary fuzzy relation, Artificial Intelligence Review 57(2024)68

2024
[4]

Baczy´ nski, B

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

2008
[5]

Baets, R

B.D. Baets, R. Mesiar,T-partitions, Fuzzy Sets and Systems 97(2)(1998)211-223

1998
[6]

Baets, R

B.D. Baets, R. Mesiar, Metrics andT-equalities, Journal of Mathematical Analysis and Applications 267(2002)531-547

2002
[7]

Bandler, L

W. Bandler, L. Kohout, Fuzzy power sets and fuzzy implication operators, Fuzzy Sets and Systems 4(1980)13-30

1980
[8]

B˘ elohl´ avek, V

R. B˘ elohl´ avek, V. Vychodil, Algebras with fuzzy equalities, Fuzzy Sets and Systems 157(2006)161-201

2006
[9]

Boixader, On the relationship betweenT-transitivity and approximate equality, Fuzzy Sets and Systems 133(2003)161-169

D. Boixader, On the relationship betweenT-transitivity and approximate equality, Fuzzy Sets and Systems 133(2003)161-169. 23

2003
[10]

Boixader, J

D. Boixader, J. Recasens, On the degree of transitivity of a fuzzy relation, Fuzzy Sets and Systems 440(2022)1-20

2022
[11]

Bustince, E

H. Bustince, E. Barrenechea and M. Pagola, Restricted equivalence functions, Fuzzy Sets and Systems 157(17)(2006)2333-2346

2006
[12]

J. L. Castro, E. Trillas, J.M. Zurita, Non-monotonic fuzzy reasoning, Fuzzy Sets and Systems 94(1998)217-225

1998
[13]

M.D. Cock, E. Kerre, On (un)suitable fuzzy relations to model approximate equality, Fuzzy Sets and Systems 133(2003)137-153

2003
[14]

Dan, B.Q

Y.X. Dan, B.Q. Hu, J.S. Qiao, Some results on the degree of symmetry of fuzzy relations, Fuzzy Sets and Systems 360(2019)1-32

2019
[15]

Demirci, L-equivalence relations on L-fuzzy sets, L-partitions of L-fuzzy sets and their one-to-one connections, International Journal of Approximate Reasoning 111(2019)21-34

M. Demirci, L-equivalence relations on L-fuzzy sets, L-partitions of L-fuzzy sets and their one-to-one connections, International Journal of Approximate Reasoning 111(2019)21-34

2019
[16]

Fodor, M

J. Fodor, M. Roubens, Fuzzy Preference Modelling and Multicriteria Decision Support, Kluwer, Dordrecht, 1994

1994
[17]

Garmendia, J

L. Garmendia, J. Recasens, How to makeT-transitive a proximity relation, IEEE Trans- action on Fuzzy Systems 17(2009)200-207

2009
[18]

Grabisch, J.L

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

2009
[19]

Grzegorzewski, Probabilistic implications, Fuzzy Sets and Systems 226(2013)53-66

P. Grzegorzewski, Probabilistic implications, Fuzzy Sets and Systems 226(2013)53-66

2013
[20]

Gupta, B

M. Gupta, B. Jayaram, Fuzzy compatibility relations and pseudo-monometrics: Some correspondences, Fuzzy Sets and Systems 451(2022)342-360

2022
[21]

Hern´ andez, I.P

M.O. Hern´ andez, I.P. Cabrera, P. Cordero et al., Fuzzy closure relations, Fuzzy Sets and Systems 450(2022)118-132

2022
[22]

Q. He, H.X. Li, Z.Z. Shi, E.S. Lee, Fuzzy clustering method based on perturbation, Computers and Mathematics with Applications 46(2003)929-946

2003
[23]

Hu, Foundations of Fuzzy Theory (2nd ed.), Publication of Wehan University, Wuhan, 2010(Chinese)

B.Q. Hu, Foundations of Fuzzy Theory (2nd ed.), Publication of Wehan University, Wuhan, 2010(Chinese)

2010
[24]

Jacas, J

J. Jacas, J. Recasens, Fixed points and generators of fuzzy relations, Journal of Mathe- matical Analysis and Applications 186(1994)21-29

1994
[25]

Jacas, J

J. Jacas, J. Recasens, FuzzyT-transitive relations: eigenvectors and generators, Fuzzy Sets and Systems 72(1995)147-154

1995
[26]

Jaurrieta, Z

M.F. Jaurrieta, Z. Tak´ a˘ c, I.R. 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. 24

2023
[27]

Klawonn, R

F. Klawonn, R. Kruse, Equality relations as a basis for fuzzy control, Fuzzy Sets and Systems 54(1993)147-156

1993
[28]

Klement, R

E.P. Klement, R. Mesiar, E. Pap, Triangular Norms, Kluwer Academic Publishers, Dor- drecht, London, 2000

2000
[29]

Li, Research on the Theory and Method of Intelligent Diagnosis Model for Vibration Faults in Steam Turbine Generator Units, Chongqing University, 1999, Doctoral disser- tation

H. Li, Research on the Theory and Method of Intelligent Diagnosis Model for Vibration Faults in Steam Turbine Generator Units, Chongqing University, 1999, Doctoral disser- tation

1999
[30]

Massanet, J

S. Massanet, J. Recasens, J. Torrens, Fuzzy implication functions based on powers of continuous t-norms, International Journal of Approximate Reasoning 83(2017)265-279

2017
[31]

Montes, I

S. Montes, I. Couso, P. Gil One-to-one correspondences betweenϵ-partitions, (1-ϵ)- equivalences andϵ-pseudometrics, Fuzzy Set and Systems 124(2001)87-95

2001
[32]

Naessens, H.D

H. Naessens, H.D. Meyer, B.D. Baets, Algorithms for the computation ofT-transitive closures, IEEE Transaction on Fuzzy Systems 10(2002)541-551

2002
[33]

Nobuhara, K

H. Nobuhara, K. Hirota, S. Sessa, W. Pedrycz, Efficient decomposition methods of fuzzy relation and their application to image decomposition, Applied Soft Computing 5(2005)399-408

2005
[34]

A.D. Nola, S. Sessa, W. Pedrycz, E. Sanchez, Fuzzy Relation Equations and Their Ap- plications to Knowledge Engineering, Norwell, MA: Kluwer, 1989

1989
[35]

Qian, X, Han, Y

J. Qian, X, Han, Y. Yu, et al., Research on multi-granularity sequential three- way decisions based on the fuzzyT-equivalence relation, Applied Soft Computing 149(2023)110980

2023
[36]

S´ anchez, P.F

T.C. S´ anchez, P.F. Parra, O. Valero, The aggregation of transitive fuzzy relations revis- ited, Fuzzy Sets and Systems 446(2022)243-260

2022
[37]

Y. Shi, B. Van Gasse, D. Ruan, E.E. Kerre, On the first place antitonicity in QL- implications, Fuzzy Sets and Systems 159(2008)2998-3013

2008
[38]

J. Sun, W. Yao,β-fuzzy equivalence relations,β-fuzzy partitions and the rough set model, Fuzzy Sets and Systems 471(2023)108670

2023
[39]

X.Q. Tang, P. Zhu, Hierarchical clustering problems and analysis of fuzzy proximity relation on granular space, IEEE Transactions on Fuzzy Systems 21(2013)814-824

2013
[40]

Valverde, On the structure ofF-indistinguishability operators, Fuzzy Sets and Systems 17(1985)313-328

L. Valverde, On the structure ofF-indistinguishability operators, Fuzzy Sets and Systems 17(1985)313-328

1985
[41]

X. Wang, Y. Xue, Traces and property indicators of fuzzy relations, Fuzzy Sets and Systems 246(2014)78-90

2014
[42]

Yager, On some new classes of implication operators and their role in approximate reasoning, Information Sciences 167(2004)193-216

R. Yager, On some new classes of implication operators and their role in approximate reasoning, Information Sciences 167(2004)193-216. 25

2004
[43]

Yang, H.M

M.S. Yang, H.M. Shih, Cluster analysis based on fuzzy relations, Fuzzy Sets and Systems 120(2001)197-212

2001
[44]

Zadeh, Similarity relations and fuzzy orderings, Information Sciences 3(1971)177- 200

L.A. Zadeh, Similarity relations and fuzzy orderings, Information Sciences 3(1971)177- 200

1971
[45]

Zadeh, The concept of a linguistic variable and its application to approximate rea- soning I, II, III, Information Sciences 8(1975)199-249, 301-357, 9(1975)43-80

L.A. Zadeh, The concept of a linguistic variable and its application to approximate rea- soning I, II, III, Information Sciences 8(1975)199-249, 301-357, 9(1975)43-80

1975
[46]

Zhang, X.Y

C.H. Zhang, X.Y. Dong, S.Z. Zeng, L.A. Carlos, Dual consistency-driven group decision making method based on fuzzy preference relation, Expert Systems With Applications 238(2024)122228. 26

2024