Open Access
Issue
Published:
01 March 2023
A Goodness-of-Fit Test Based on Fuzzy Random Variables
)
Download citation
726
Views
82
Downloads
0
Crossref
0
WoS
0
Scopus
N/A
CSCD
During the last decades, several methods have been proposed for Kolmogorov−Smirnov one-sample test based on fuzzy random variables to describe the impression of classical random variables. However, such techniques do not discuss the modeling of imprecise observations and simulation of such data from the distribution of a fuzzy random variable. Moreover, such methods rely on a fuzzy cumulative distribution function with known parameters. In this paper, however, a modified Kolmogorov−Smirnov one-sample test is introduced based on a novel notion of fuzzy random variables which comes down to model fuzziness and randomness in the distribution of population in a frequently used family of probability distributions called location and scale distribution functions. A method of moment estimator was also utilized to estimate the location and scale parameters. Then, a notion of non-fuzzy Kolmogorov−Smirnov one-sample test was developed based on fuzzy hypotheses. Monte Carlo simulation was also employed to evaluate the critical value corresponding to a significance level and the performance of the test using power studies. Comparing the observed test statistics and the given fuzzy significance level, a classical procedure was finally used to accept or reject the null fuzzy hypothesis. Two numerical examples including a simulation study and an applied example were provided to clarify the discussions in this paper. The proposed method was also compared with some existing methods. The goodness-of-fit results demonstrated that the proposed Kolmogorov−Smirnov provides an efficient tool to handle statistical inference fuzzy observations.
menu
A Goodness-of-Fit Test Based on Fuzzy Random Variables
)
Abstract
During the last decades, several methods have been proposed for Kolmogorov−Smirnov one-sample test based on fuzzy random variables to describe the impression of classical random variables. However, such techniques do not discuss the modeling of imprecise observations and simulation of such data from the distribution of a fuzzy random variable. Moreover, such methods rely on a fuzzy cumulative distribution function with known parameters. In this paper, however, a modified Kolmogorov−Smirnov one-sample test is introduced based on a novel notion of fuzzy random variables which comes down to model fuzziness and randomness in the distribution of population in a frequently used family of probability distributions called location and scale distribution functions. A method of moment estimator was also utilized to estimate the location and scale parameters. Then, a notion of non-fuzzy Kolmogorov−Smirnov one-sample test was developed based on fuzzy hypotheses. Monte Carlo simulation was also employed to evaluate the critical value corresponding to a significance level and the performance of the test using power studies. Comparing the observed test statistics and the given fuzzy significance level, a classical procedure was finally used to accept or reject the null fuzzy hypothesis. Two numerical examples including a simulation study and an applied example were provided to clarify the discussions in this paper. The proposed method was also compared with some existing methods. The goodness-of-fit results demonstrated that the proposed Kolmogorov−Smirnov provides an efficient tool to handle statistical inference fuzzy observations.
References(25)
P. Grzegorzewski and H. Szymanowski, Goodness-of-fit tests for fuzzy data, Inf. Sci., vol. 288, pp. 374–386, 2014.
S. R. Eliason and R. Stryker, Goodness-of-fit tests and descriptive measures in fuzzy-set analysis, Sociol. Methods Res., vol. 38, no. 1, pp. 102–146, 2009.
G. Hesamian, and J. Chachi, Two-sample Kolmogorov–Smirnov fuzzy test for fuzzy random variables, Stat. Pap., vol. 56, no. 1, pp. 61–82, 2015.
G. Hesamian and S. M. Taheri, Fuzzy empirical distribution function: Properties and application, Kybernetika, vol. 49, no. 6, pp. 962–982, 2013.
I. Couso and D. Dubois, On the variability of the concept of variance for fuzzy random variables, IEEE Trans. Fuzzy Syst., vol. 17, no. 5, pp. 1070–1080, 2009.
R. Feron, Ensembles aléatoires flous, (in French), C. R. Acad. Sci. Paris, vol. 282, pp. 903–906, 1976.
M. L. Puri and D. A. Ralescu, Fuzzy random variables, J. Math. Anal. Appl., vol. 114, no. 2, pp. 409–422, 1986.
H. Kwakernaak, Fuzzy random variables—I. Definitions and theorems, Inf. Sci., vol. 15, no. 1, pp. 1–29, 1978.
H. Kwakernaak, Fuzzy random variables—II. Algorithms and examples for the discrete case, Inf. Sci., vol. 17, no. 3, pp. 23–278, 1979.
Y. K. Liu and B. Liu, Fuzzy random variables: A scalar expected value operator, Fuzzy Optim. Decis. Mak., vol. 2, pp. 143–160, 2003.
V. Krätschmer, A unified approach to fuzzy random variables, Fuzzy Sets Syst., vol. 123, no. 1, pp. 1–9, 2001.
A. F. Shapiro, Fuzzy random variables, Insur.: Math. Econ., vol. 44, no. 2, pp. 307–314, 2009.
M. Á. Gil, M. López-Díaz, and D. A. Ralescu, Overview on the development of fuzzy random variables, Fuzzy Sets Syst., vol. 157, no. 19, pp. 2546–2557, 2006.
A. Ban, A. Brândaş, L., Coroianu, C. Negruţiu, and O. Nica, Approximations of fuzzy numbers by trapezoidal fuzzy numbers preserving the ambiguity and value, Comput. Math. Appl., vol. 61, no. 5, pp. 1379–1401, 2011.
G. Hesamian and M. Shams, Parametric testing statistical hypotheses for fuzzy random variables, Soft Comput., vol. 20, no. 4, pp. 1537–1548, 2016.
M. G. Akbari and G. Hesamian, Linear model with exact inputs and interval-valued fuzzy outputs, IEEE Trans. Fuzzy Syst., vol. 26, no. 2, pp. 518–530, 2018.
G. Hesamian and M. G. Akbari, Fuzzy cumulative entropy and its estimation based on fuzzy random variables, Int. J. Syst. Sci., vol. 53, no. 5, pp. 982–991, 2022.
Publication history
Copyright
Rights and permissions
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
FEEDBACK 
TOP