Abstract
Uncertainty measurements underlie the system interaction and data learning. Their relevant studies are extensive for the single-valued decision systems, but become relatively less for the interval-valued decision systems. Thus, three-level and three-way uncertainty measurements of the interval-valued decision systems are proposed, mainly by systematically constructing vertical-horizontal weighted entropies. Firstly, the interval-valued decision systems are endowed with three-level structures, including Micro-Bottom, Meso-Middle, and Macro-Top. Secondly, a three-level decomposition is hierarchically made for the existing conditional entropy. Thirdly, three-way weighted entropies are systematically and hierarchically constructed at the three levels, and they achieve their hierarchy, systematicness, algorithm, boundedness, and granulation monotonicity/non-monotonicity. The three-level and three-way weighted entropies deepen and extend the conditional entropy, and they realize the ingenious criss-cross informatization for the interval-valued decision systems. Their effectiveness of uncertainty measurements is ultimately verified by table examples and data experiments.
Similar content being viewed by others
References
Pawlak Z (1991) Rough sets: theoretical aspects of reasoning about data. Kluwer Academic Publishers, Dordrecht
Li JH, Huang CC, Qi JJ, Qian YH, Liu WQ (2017) Three-way cognitive concept learning via multi-granularity. Inf Sci 378:244–263
Saha I, Sarkar JP, Maulik U (2019) Integrated rough fuzzy clustering for categorical data analysis. Fuzzy Sets Syst 361:1–32
Wang ZH, Feng QR, Wang H (2019) The lattice and matroid representations of definable sets in generalized rough sets based on relations. Inf Sci 485:505–520
Zhang XY, Gou HY, Lv ZY, Miao DQ (2021) Double-quantitativedistance measurement and classification learning based on the tri-level granular structure of neighborhood system. Knowl Based Syst. https://doi.org/10.1016/j.knosys.2021.106799
Zhan JM, Wang QM (2019) Certain types of soft coverings based rough sets with applications. Int J Mach Learn Cybern 10(5):1065–1076
Prasad M, Tripathi S, Dahal K (2020) An efficient feature selection based Bayesian and rough set approach for intrusion detection. Appl Soft Comput 87:105980
Xu WH, Yu JH (2017) A novel approach to information fusion in multi-source datasets: a granular computing viewpoint. Inf Sci 378:410–423
Jia XY, Rao Y, Shang L, Li TJ (2020) Similarity-based attribute reduction in rough set theory: a clustering perspective. Int J Mach Learn Cybern 11:1047–1060
Yang XB, Liang SC, Yu HL, Gao S, Qian YH (2019) Pseudo-label neighborhood rough set: measures and attribute reductions. Int J Approx Reason 105:112–129
Li WW, Jia XY, Wang L, Zhou B (2019) Multi-objective attribute reduction in three-way decision-theoretic rough set model. Int J Approx Reason 105:327–341
Sun L, Zhang XY, Qian YH, Xu JC, Zhang SG (2019) Feature selection using neighborhood entropy-based uncertainty measures for gene expression data classification. Inform Sci 502:18–41
Wang CZ, Huang Y, Shao MW, Fan XD (2019) Fuzzy rough set-based attribute reduction using distance measures. Knowl Based Syst 164:205–212
Liu KY, Yang XB, Fujita HM, Liu D, Yang X, Qian YH (2019) An efficient selector for multi-granularity attribute reduction. Inform Sci 505:457–472
Zhang QH, Yang SH, Wang GY (2017) Measuring uncertainty of probabilistic rough set model from its three regions. IEEE Trans Syst Man Cybern 47(12):3299–3309
Wang GY, Ma XA, Yu H (2015) Monotonic uncertainty measures for attribute reduction in probabilistic rough set model. Int J Approx Reason 59:41–67
Gao C, Lai ZH, Zhou J, Wen JJ, Wong WK (2019) Granular maximum decision entropy-based monotonic uncertainty measure for attribute reduction. Int J Approx Reason 104:9–24
Qian J, Dang CY, Yue XD, Zhang N (2017) Attribute reduction for sequential three-way decisions under dynamic granulation. Int J Approx Reason 85:196–216
Hu J, Li TR, Wang HJ, Fujita HM (2016) Hierarchical cluster ensemble model based on knowledge granulation. Knowl Based Syst 91:179–188
Wang GY, Zhao J, An JJ, Wu Y (2005) A comparative study of algebra viewpoint and information viewpoint in attribute reduction. Fund Inform 68(3):289–301
Shannon C (1948) The mathematical theory of communiccation. Bell Syst Tech J 27:379–423
Wang ZH, Yue HF, Deng JP (2019) An uncertainty measure based on lower and upper approximations for generalized rough set models. Fund Inform 166(3):273–296
Zhao JY, Zhang ZL, Han CZ, Zhou ZF (2015) Complement information entropy for uncertainty measure in fuzzy rough set and its applications. Soft Comput 19(7):1997–2010
Chen YM, Xue Y, Ma Y, Xu FF (2017) Measures of uncertainty for neighborhood rough sets. Knowl Based Syst 120:226–235
Zhang XY, Miao DQ (2017) Three-layer granular structures and three-way informational measures of a decision table. Inform Sci 412:67–86
Zhang XY, Yao H, Lv ZY, Miao DQ (2021) Class-specific information measures and attribute reducts for hierarchy and systematicness. Inform Sci. https://doi.org/10.1016/j.ins.2021.01.080
Tang LY, Zhang XY, Mo ZW (2020) A weighted complement-entropy system based on tri-level granular structures. Int J Gen Syst. https://doi.org/10.1080/03081079.2020.1806833
Dai JH, Yan YJ, Li ZW, Liao BS (2018) Dominance-based fuzzy rough set approach for incomplete interval-valued data. J Intell Fuzzy Syst 34(1):423–436
Lin BY, Xu WH (2018) Multi-granulation rough set for incomplete interval-valued decision information systems based on multi-threshold tolerance relation. Symmetry 10(6):208
Guo YT, Tsang ECC, Xu WH, Chen DG (2020) Adaptive weighted generalized multi-granulation interval-valued decision-theoretic rough sets. Knowl Based Syst 187:104804
Dai JH, Wang WT, Xu Q, Tian HW (2012) Uncertainty measurement for interval-valued decision systems based on extended conditional entropy. Knowl Based Syst 27:443–450
Dai JH, Wang WT, Mi JS (2013) Uncertainty measurement for interval-valued information systems. Inform Sci 251:63–78
Dai JH, Wei BJ, Zhang XH, Zhang QL (2017) Uncertainty measurement for incomplete interval-valued information systems based on \(\alpha \)-weak similarity. Knowl Based Syst 136:159–171
Xie NG, Liu M, Li A, Zhang GQ (2019) New measures of uncertainty for an interval-valued information system. Inform Sci 470:156–174
Yao YY (2016) Three-way decisions and cognitive computing. Cogn Comput 8(4):543–554
Yao YY (2020) Tri-level thinking: models of three-way decision. Int J Mach Learn Cybern 11:947–959
Yang B, Li JH (2020) Complex network analysis of three-way decision researches. Int J Mach Cybern 11:973–987
Zhang XY, Tang X, Yang JL, Lv ZY (2020) Quantitative three-way class-specific attribute reducts based on region preservations. Int J Approx Reason 117:96–121
Zhang XY, Miao DQ (2017) Three-way attribute reducts. Int J Approx Reason 88:401–434
Lang GM, Miao DQ, Fujita H (2020) Three-way group conflict analysis based on Pythagorean fuzzy set theory. IEEE Trans Fuzzy Syst 28(3):447–461
Zhao XR, Hu BQ (2020) Three-way decisions with decision-theoretic rough sets in multiset-valued information tables. Inform Sci 507:684–699
Sun BZ, Chen XT, Zhang LY, Ma WM (2020) Three-way decision making approach to conflict analysis and resolution using probabilistic rough set over two universes. Inform Sci 507:809–822
Hu MJ, Yao YY (2019) Structured approximations as a basis for three-way decisions in rough set theory. Knowl Based Syst 165:92–109
Yao YY (2018) Three-way decision and granular computing. Int J Approx Reason 103:107–123
Yao YY (2020) Three-way granular computing, rough sets, and formal concept analysis. Int J Approx Reason 116:106–125
Mu TP, Zhang XY, Mo ZW (2019) Double-granule conditional-entropies based on three-level granular structures. Entropy 21(7):657
Nakahara Y, Sasaki M, Gen M (1992) On the linear programming problems with interval coefficients. Comput Ind Eng 23(1–4):301–304
Nakahara Y (1998) User oriented ranking criteria and its application to fuzzy mathematical programming problems. Fuzzy Sets Syst 94(3):275–286
Ishibuchi H, Tanaka H (1988) Formulation and analysis of linear programming problem with interval coefficients. J Jpn Ind Manag Assoc 40:320–329 (in Japanese)
Acknowledgements
The authors thank both the editors and the reviewers for their valuable suggestions, which substantially improve this paper. This work was supported by National Natural Science Foundation of China (61673285 and 11671284), Sichuan Science and Technology Program of China (2021YJ0085 and 2019YJ0529), and A Joint Research Project of Laurent Mathematics Center of Sichuan Normal University and National-Local Joint Engineering Laboratory of System Credibility Automatic Verification.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix A. Proof of Proposition 2
Proof
Regarding \(H_{lklhd}(D_{j}/{S_B^\theta (u_i)})\) and by Lemma 2, we have \(0\le -\frac{1}{n}P({S_B^\theta (u_i)})P({D_{j}/S_B^\theta (u_i)})\log _{2}P({D_{j}/S_B^\theta (u_i)})\le \frac{1}{n}\times 1\times \frac{1}{e ln2}\). Next prove the relevant realizability. Possible value \(P({D_{j}/S_B^\theta (u_i)})=0\) guides \(H_{lklhd}(D_{j}/{S_B^\theta (u_i)})=0\). On the other hand, possible value \(P({S_B^\theta (u_i)})=1\) and approximate value \(P({D_{j}/S_B^\theta (u_i)})\rightarrow \frac{1}{e}\) can realize \(H_{lklhd}(D_{j}/{S_B^\theta (u_i)})\rightarrow \frac{1}{n\times eln2}\), and the relevant case corresponds to \({S_B^\theta (u_i)}=U\), \(|D_{j}|\rightarrow \frac{n}{e}\). The other measures \(H_{prior}({S_B^\theta (u_i)})_{D_{j}}\) and \(H_{pstrr}({S_B^\theta (u_i)}/D_{j})\) have the similar proof. \(\square \)
Appendix B. Proof of Lemma 3
Proof
(1) This proof mainly comes from Ref. [31]. Let \(z=\frac{x}{{x+y}}\in (0,1]\) in f(x, y), and then we have
(2) g(x, y) has partial derivatives:
Thus, we concern only \(G(x,y)>0\), \(G(x,y)=0\), \(G(x,y)<0\) to present \(\frac{\partial {g(x,y)}}{\partial {x}}\), and fixed \(y>0\) is a premise. For this purpose, suppose \(g_1(x,y)=\frac{1}{ln2}(\frac{y}{x}+1)\) and \(g_2(x,y)=\log _2\frac{x+y}{n}\) to offer \(G(x,y)=g_1(x,y)-g_2(x,y)\), and \(G(x,y)=0\) (i.e., Eq. (36)) becomes a basic equation to solve x. Intuitively, the decreasing hyperbolic-curve \(g_1(x,y)\) and the increasing logarithmic-function \(g_2(x,y)\) have a sole intersection point with horizontal ordinate \(x^\dag \), so \(G(x,y)\ge 0\) and \(G(x,y)\le 0\) are accordingly realized by \(x\le x^\dag \) and \(x\ge x^\dag \). More deeply,
because of G(x, y)’s x-continuity and \(\frac{\partial {G(x,y)}}{\partial {x}}=\frac{\partial ^2 {g(x,y)}}{\partial {x}^2}<0\), key equation \(G(x,y)=0\) (i.e., Eq. (36)) has a unique zero-point \(x^\dag \) (i.e., the sole x-univariate stationary point of g(x, y)), while \(x\le x^\dag \) and \(x\ge x^\dag \) respectively correspond to \(G(x,y)\ge 0\) and \(G(x,y)\le 0\). In a word, \(\frac{\partial {g(x,y)}}{\partial {x}}=G(x,y)\) acquires the required positive-negative identification by the sole root \(x^\dag \in (0,+\infty )\). \(\square \)
Appendix C. Three experimental Interval-VDSs from Ref. [31]: Tables 7, 8, 9
Rights and permissions
About this article
Cite this article
Liao, S., Zhang, X. & Mo, Z. Three-level and three-way uncertainty measurements for interval-valued decision systems. Int. J. Mach. Learn. & Cyber. 12, 1459–1481 (2021). https://doi.org/10.1007/s13042-020-01247-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13042-020-01247-8