On a new PGDUS transformed model using Inverse Weibull distribution

2025

Read On a new PGDUS transformed model using Inverse Weibull distribution (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-12307.txt.

Introduction

Gauthami and Chacko build a new chance spread by use of the Power Generalized DUS shift on the inverse Weibull spread. The shift counts in work on trust of gear, most for twin paths side by side. The inverse Weibull model is much used for life span data. The team works out the base chance traits of the new model.

Traits not known are gauged by best chance and by best yield of gaps. A mock run checks the gauges, and the trust trait that weighs two life spans is looked at. A fit to a real world data set shows the new model beats rival spreads. The team sets it out as a choice with give for work on trust and life span.

Catalog record

YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE42 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-12307

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

On a new PGDUS transformed model using Inverse Weibull distribution Abstract The Power Generalized DUS (PGDUS) Transformation is significant in reliability theory, especially for analyzing parallel systems. From the Generalized Extreme Value distribution, Inverse Weibull model particularly has wide applicability in statistics and reliability theory. In this paper we consider PGDUS transformation of Inverse Weibull distribution. The basic statistical characteristics of the new model are derived, and unknown parameters are estimated using Maximum likelihood and Maximum product of spacings methods. Simulation analysis and the reliability parameter P ⁡ ( T 2 < T 1 ) P(T_{2}<T_{1}) are explored. The effectiveness of the model in fitting a real-world dataset is demonstrated, showing better performance compared to other competing distributions. Gauthami P. & Chacko V. M. Department of Statistics St. Thomas College (Autonomous), Thrissur, University of Calicut, Kerala, India - 680001 [email protected] , [email protected] Keywords: PGDUS transformation; Inverse Weibull distribution; Maximum likelihood estimation; Maximum product of spacings estimation. 1. INTRODUCTION To model reliability and lifetime data, a variety of distributions have been developed in the fields of statistics and reliability engineering. If the current distributions are less effective in some circumstances, researchers will inevitably have to find alternative distributions that fit the provided data. Also, it is preferable to have a suitable lifetime model when thinking about parallel systems, in which each of the components is known to have some well-known distributions. PGDUS transformation is a method for modeling parallel systems where each of the components has its distribution DUS transformed to any baseline model, which makes the transformation method more applicable in the field of reliability modeling. A model selection using DUS transformation would be more appropriate because it is a concise method for creating more appropriate distributions without adding more parameters and to fit suitable models for the given data. Focusing on the basics of transformation techniques, sometimes, it would be better to introduce new distributions by incorporating a shape parameter, since shape of the distribution always changes dynamically from situation to situation. Let T be a random variable, consider the transformation with new Cumulative Distribution Function (CDF) defined by F_{T_{new}}(t)=(F_{T_{base}}(t))^{\gamma} where, F_{T_{base}}(t) represents the CDF of a baseline distribution and \gamma>0 is an added shape parameter that enriches the model. Kumar et al. 2015 developed the previously mentioned DUS transformation without increasing the number of parameters and introduced the DUS Exponential (DUS-E) distribution, where this approach has proven effective in transforming baseline distributions to better reflect the complexities of real-world data. A number of studies have further explored the DUS transformation in the literature. Using both traditional and Bayesian methods, Tripathi et al. 2021 investigated parameter estimation for the DUS-E distribution with an emphasis on upper record values. By applying the DUS transformation to the Lomax distribution, Deepthi and Chacko 2020 created the DUS Lomax (DUS-L) model and investigated its use in stress-strength reliability. Another significant advancement was the work of Gauthami and Chacko 2021 , who proposed and thoroughly examined the DUS Inverse Weibull (DUS-IW) distribution, which is distinguished by its upside-down bathtub-shaped hazard rate. Related studies that expanded these findings to the DUS-Kumaraswamy (DUS-K) distribution include those by Karakaya et al. 2021 and Anakha and Chacko 2021 . The applications of these transformations have been expanded by further studies by Kavya and Manoharan 2020 , Nayana et al. 2022 , Khan and Mustafa 2023 , and Hassan and Aslam 2023 , which have used them in a variety of datasets and further established their use in reliability modeling. Modeling and Inference are fundamental in statistical research, especially in developing lifetime distributions tailored to component reliability. Many lifetime models have emerged through diverse transformation techniques, which allow for the creation of component-specific lifetime distributions. However, a more comprehensive objective lies in modeling the reliability of an entire system, especially systems with parallel-connected components. In this vein, the PGDUS transformation, introduced by Thomas and Chacko 2021 , provides a valuable framework, enabling the modeling of an entire system when each component follows a DUS-transformed model. This utility of the PGDUS transformation is showcased in their proposed PGDUS-Exponential (PGDUS-E) distribution. Emphasizing this transformation, which effectively extends the DUS model through exponentiation, recent work has expanded this approach to the Weibull (PGDUS-W) and Lomax (PGDUS-L) distributions (see Thomas and Chacko 2023 ) and the Inverse Kumaraswamy (PGDUS-IK) (see Amrutha and Chacko 2024 ) distribution. If T is the random variable under consideration, following the baseline distribution with CDF, F_{T_{base}}(t) and Probability Density Function (PDF), f_{T_{base}}(t) , then the CDF of the PGDUS distribution is \displaystyle F_{T_{PGDUS}}(t)=\begin{cases}\left(\frac{e^{F_{T_{base}}(t)}-1}{e-1}\right)^{\gamma}&;\,t>0,\,\gamma>0\\ 0&;\,Otherwise\end{cases}. (1.1) The PDF of the PGDUS distribution is \displaystyle f_{T_{PGDUS}}(t)=\begin{cases}\frac{\gamma}{(e-1)^{\gamma}}\hskip 5.69054ptf_{T_{base}}(t)\,e^{F_{T_{base}}(t)}\left(e^{F_{T_{base}}(t)}-1\right)^{\gamma-1}&;\,t>0,\,\gamma>0\\ 0&;\,Otherwise\end{cases}. (1.2) The survival function of the PGDUS distribution is \displaystyle S_{T_{PGDUS}}(t)=\begin{cases}1-\left(\frac{e^{F_{T_{base}}(t)}-1}{e-1}\right)^{\gamma}&;\,t>0,\,\gamma>0\\ 0&;\,Otherwise\end{cases}. (1.3) The failure-rate function of the PGDUS distribution is \displaystyle h_{T_{PGDUS}}(t)=\begin{cases}\frac{\gamma\,f_{T_{base}}(t)\,e^{F_{T_{base}}(t)}\left(e^{F_{T_{base}}(t)}-1\right)^{\gamma-1}}{(e-1)^{\gamma}-\left(e^{F_{T_{base}}(t)}-1\right)^{\gamma}}&;\,t>0,\,\gamma>0\\ 0&;\,Otherwise\end{cases}. (1.4) In this paper, we explore the PGDUS transformation with a particular emphasis on selecting an appropriate baseline model. The baseline model selection problem is quite complicated. According to the literature, the field of statistics highly values the Generalized Extreme Value (GEV) distribution family, which includes the Gumbel, Weibull, and Inverse Weibull (IW) distributions, for its versatility in reliability theory, lifetime and survival data analysis. Originally developed by Waloddi Weibull in the 1930s to characterize material strength, the Weibull distribution is still often used in statistical analysis of survival and reliability data. In their study of the three-parameter Weibull distribution, Zanakis and Kyparisis 1986 analyzed its estimation procedures and pointed out that it can be difficult to estimate all three parameters at once, which frequently causes researchers to rely on the more straight-forward two-parameter Weibull distribution. Similarly, constraints have been identified in the Maximum Likelihood (ML) estimation for the three-parameter Weibull model in early researches (see Harter 1970 ; Harter 1971 ). Due in large part to its adaptability in modeling a variety of experimental data, subsequent research, including as studies by Mudholkar et al. 1996 and Evans et al. 2019 , has made a substantial contribution to the understanding of the Weibull distribution. This distribution transforms into the IW distribution, sometimes referred to as the Fréchet (F) distribution, when inverse transformation is applied, and is characterized by its three-parameter CDF, ( 1.5 ) and PDF, ( 1.6 ). \displaystyle F_{T_{F}}(t)=\begin{cases}e^{-\left(\frac{t-\alpha}{\theta}\right)^{-\lambda}}&;\,t>0,\,\lambda>0,\,\theta>0,\alpha\in\mathbb{R}\\ 0&;\,Otherwise\end{cases}. (1.5) \displaystyle f_{T_{F}}(t)=\begin{cases}\frac{\lambda}{\theta}\,\left(\frac{t-\alpha}{\theta}\right)^{-(\lambda+1)}\,e^{-\left(\frac{t-\alpha}{\theta}\right)^{-\lambda}}&;\,t>0,\,\lambda>0,\,\theta>0,\alpha\in\mathbb{R}\\ 0&;\,Otherwise\end{cases}. (1.6) This is simply a two-parameter IW model when the third parameter \alpha equals zero, which considered to be our baseline distribution. With regard to its flexibility in expressing data with monotonic failure rates, the IW distribution is a flexible lifetime model that is widely used in domains such as engineering, survival analysis, reliability theory, and risk modeling. The IW model has proven useful for practical applications, much like the Weibull distribution. Keller et al. 1982 , Calabria and Pulcini 1989 and Calabria and Pulcini 1994 are the early research works on IW distribution. Mahmoud et al. 2003 examined order statistics and related inference for the IW model, which is well-documented in statistical literature and goes by several names, such as complementary Weibull, reciprocal Weibull, and reverse Weibull (see Drapella 1993 , Mudholkar and Kollia 1994 , and Murthy et al. 2004 ). Studies by Khan et al. 2008 and De Gusmao et al. 2011 further emphasize the many uses of the

Read the complete text