Adaptive estimation: Fuzzy data-driven gamma distribution via Bayesian and maximum likelihood approaches

Abbarapu Ashok; Nadiminti Nagamani; Abbarapu Ashok; Nadiminti Nagamani

doi:10.3934/math.2025021

AIMS Mathematics

2025, Volume 10, Issue 1: 438-459. doi: 10.3934/math.2025021

Previous Article Next Article

Research article

Adaptive estimation: Fuzzy data-driven gamma distribution via Bayesian and maximum likelihood approaches

Abbarapu Ashok ,
Nadiminti Nagamani ^,

Department of Mathematics, School of Advanced Sciences, VIT-AP University, Amaravati, Andhra Pradesh, India

Received: 06 November 2024 Revised: 10 December 2024 Accepted: 27 December 2024 Published: 09 January 2025
MSC : 62F86, 62F15

Integrating fuzzy concepts into statistical estimation offers considerable advantages by enhancing both the accuracy and reliability of parameter estimations, irrespective of the sample size and technique used. This study specifically examined the improvement of parameter estimation accuracy when dealing with fuzzy data, with a focus on the gamma distribution. We explored and evaluated a variety of estimation techniques for determining the scale parameter $ \eta $ and shape parameter $ \rho $ of the gamma distribution, employing both maximum likelihood (ML) and Bayesian methods. In the case of ML estimates, the expectation-maximization (EM) algorithm and the Newton-Raphson (NR) method were applied, with confidence intervals constructed using the Fisher information matrix. Additionally, the highest posterior density (HPD) intervals were derived through Gibbs sampling. For Bayesian estimates, the Tierney and Kadane (TK) approximation and Gibbs sampling were used to enhance the estimation process. A thorough performance comparison was undertaken using a simulated fuzzy dataset of the lifetimes of rechargeable batteries to assess the effectiveness of these methods. The methods were evaluated by comparing the estimated parameters to their true values using mean squared error (MSE) as a metric. Our findings demonstrate that the Bayesian approach, particularly when combined with the TK method, consistently produces more accurate and reliable parameter estimates compared to traditional methods. These results underscore the potential of Bayesian techniques in addressing fuzzy data and enhancing precision in statistical analyses.
- EM algorithm,
- fuzzy data,
- Gibbs sampling,
- gamma distribution,
- TK approximation
Citation: Abbarapu Ashok, Nadiminti Nagamani. Adaptive estimation: Fuzzy data-driven gamma distribution via Bayesian and maximum likelihood approaches[J]. AIMS Mathematics, 2025, 10(1): 438-459. doi: 10.3934/math.2025021

Related Papers:

Abstract

Integrating fuzzy concepts into statistical estimation offers considerable advantages by enhancing both the accuracy and reliability of parameter estimations, irrespective of the sample size and technique used. This study specifically examined the improvement of parameter estimation accuracy when dealing with fuzzy data, with a focus on the gamma distribution. We explored and evaluated a variety of estimation techniques for determining the scale parameter $ \eta $ and shape parameter $ \rho $ of the gamma distribution, employing both maximum likelihood (ML) and Bayesian methods. In the case of ML estimates, the expectation-maximization (EM) algorithm and the Newton-Raphson (NR) method were applied, with confidence intervals constructed using the Fisher information matrix. Additionally, the highest posterior density (HPD) intervals were derived through Gibbs sampling. For Bayesian estimates, the Tierney and Kadane (TK) approximation and Gibbs sampling were used to enhance the estimation process. A thorough performance comparison was undertaken using a simulated fuzzy dataset of the lifetimes of rechargeable batteries to assess the effectiveness of these methods. The methods were evaluated by comparing the estimated parameters to their true values using mean squared error (MSE) as a metric. Our findings demonstrate that the Bayesian approach, particularly when combined with the TK method, consistently produces more accurate and reliable parameter estimates compared to traditional methods. These results underscore the potential of Bayesian techniques in addressing fuzzy data and enhancing precision in statistical analyses.

References

[1]	M. Engelhardt, L. J. Bain, Uniformly most powerful unbiased tests on the scale parameter of a gamma distribution with a nuisance shape parameter, Technometrics, 19 (1977), 77–81. https://doi.org/10.1080/00401706.1977.10489502 doi: 10.1080/00401706.1977.10489502
[2]	R. E. Glaser, The ratio of the geometric mean to the arithmetic mean for a random sample from a gamma distribution, J. Amer. Stat. Assoc., 71 (1976), 480–487.
[3]	H. Aksoy, Use of gamma distribution in hydrological analysis, Turk. J. Eng. Environmental Sci., 24 (2000), 419–428.
[4]	P. A. M. Alemán, E. García, The variability of rainfall in Mexico and its determination by means of the gamma distribution, Geogr. Ann.: Ser. A Phys. Geogr, 63 (1981), 1–10. https://doi.org/10.1080/04353676.1981.11880012 doi: 10.1080/04353676.1981.11880012
[5]	R. D. Gupta, D. Kundu, Generalized exponential distribution: different methods of estimations, J. Stat. Comput. Simul., 69 (2001), 315–337. https://doi.org/10.1080/00949650108812098 doi: 10.1080/00949650108812098
[6]	S. Nadarajah, A. K. Gupta, The exponentiated gamma distribution with application to drought data, Calcutta Stat. Assoc. Bull., 59 (2007), 29–54.
[7]	S. C. Choi, R. Wette, Maximum likelihood estimation of the parameters of the gamma distribution and their bias, Technometrics, 11 (1969), 683–690.
[8]	G. Robert, Estimation of the parameters of the gamma distribution by means of a maximal invariant, Commun. Stat.-Theory Meth., 10 (1981), 1095–1110. https://doi.org/10.1080/03610928108828096 doi: 10.1080/03610928108828096
[9]	Y. S. Son, M. Oh, Bayesian estimation of the two-parameter Gamma distribution, Commun. Stat.-Simul. Comput., 35 (2006), 285–293. https://doi.org/10.1080/03610910600591925 doi: 10.1080/03610910600591925
[10]	B. Pradhan, D. Kundu, Bayes estimation and prediction of the two-parameter gamma distribution, J. Stat. Comput. Simul., 81 (2011), 1187–1198. https://doi.org/10.1080/00949651003796335 doi: 10.1080/00949651003796335
[11]	W. Chen, C. Long, R Yang, D. Yao, Maximum likelihood estimator of the location parameter under moving extremes ranked set sampling design, Acta Math. Appl. Sin. Engl. Ser., 37 (2021), 101–108. https://doi.org/10.1007/s10255-021-0998-8 doi: 10.1007/s10255-021-0998-8
[12]	D. Meng, S. Yang, H. Yang, A. M. P. De Jesus, J. Correia, S. P. Zhu, Intelligent-inspired framework for fatigue reliability evaluation of offshore wind turbine support structures under hybrid uncertainty, Ocean Eng., 307 (2024), 118213. https://doi.org/10.1016/j.oceaneng.2024.118213 doi: 10.1016/j.oceaneng.2024.118213
[13]	H. Z. Huang, M. J. Zuo, Z. Q. Sun, Bayesian reliability analysis for fuzzy lifetime data, Fuzzy Sets Syst., 157 (2006), 1674–1686. https://doi.org/10.1016/j.fss.2005.11.009 doi: 10.1016/j.fss.2005.11.009
[14]	R. Coppi, M. A. Gil, H. A. L. Kiers, The fuzzy approach to statistical analysis, Comput. Stat. Data Anal., 51 (2006), 1–14. https://doi.org/10.1016/j.csda.2006.05.012 doi: 10.1016/j.csda.2006.05.012
[15]	T. Denoeux, Maximum likelihood estimation from fuzzy data using the EM algorithm, Fuzzy Sets Syst., 183 (2011), 72–91. https://doi.org/10.1016/j.fss.2011.05.022 doi: 10.1016/j.fss.2011.05.022
[16]	A. Pak, G. A. Parham, M. Saraj, On estimation of Rayleigh scale parameter under doubly type-Ⅱ censoring from imprecise data, J. Data Sci., 11 (2013), 305–322.
[17]	A. Pak, G. A. Parham, M. Saraj, Inference for the Weibull distribution based on fuzzy data, Rev. Colomb. Estad., 36 (2013), 337–356.
[18]	N. B. Khoolenjani, F. Shahsanaie, Estimating the parameter of Exponential distribution under Type-Ⅱ censoring from fuzzy data, J. Stat. Theory Appl., 15 (2016), 181–195. https://doi.org/10.2991/jsta.2016.15.2.8 doi: 10.2991/jsta.2016.15.2.8
[19]	K. S. Kula, T. E. Dallkilic, Parameter estimation for Pareto distribution and Type-Ⅱ fuzzy logic, Gazi Uni. J. Sci., 30 (2017), 251–258.
[20]	M. S. Yang, C. F. Su, On parameter estimation for normal mixtures based on fuzzy clustering algorithms, Fuzzy Sets Syst., 68 (1994), 13–28. https://doi.org/10.1016/0165-0114(94)90270-4 doi: 10.1016/0165-0114(94)90270-4
[21]	I. Gath, A. B. Geva, Fuzzy clustering for the estimation of the parameters of the components of mixtures of normal distributions, Pattern Recogn. Lett., 9 (1989), 77–86. https://doi.org/10.1016/0167-8655(89)90040-8 doi: 10.1016/0167-8655(89)90040-8
[22]	H. Basharat, S. Mustafa, S. Mahmood, Y. B. Jun, Inference for the linear combination of two independent exponential random variables based on fuzzy data, Hacet. J. Math. Stat., 48 (2019), 1859–1869. https://doi.org/10.15672/hujms.470452 doi: 10.15672/hujms.470452
[23]	L. Tierney, J. B. Kadane, Accurate approximations for posterior moments and marginal densities, J. Amer. Stat. Assoc., 81 (1986), 82–86.
[24]	L. Devroye, A simple algorithm for generating random variates with a log-concave density, Computing, 33 (1984), 247–257.

Reader Comments

Your name:*

Email:*
© 2025 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)