NOVA DISTRIBUIÇÃO WEIBULL INVERTIDA COSSENO TIPO I: METODOLOGIA E ESTIMAÇÃO
NOVA DISTRIBUIÇÃO WEIBULL INVERTIDA COSSENO TIPO I: METODOLOGIA E ESTIMAÇÃO
-
DOI: 10.22533/at.ed.0952322125
-
Palavras-chave: Gerador TIC, distribuição Weibull Invertida, estimação de máxima verossimilhança, simulações de Monte Carlo
-
Keywords: TIC generator, Inverted Weibull distribution, maximum likelihood estimation, Monte Carlo simulation
-
Abstract: The development and improvement of statistical methodologies, combined with the efficient use of databases, enable the extraction of valuable information to support decision-making and uncover hidden patterns and behaviors, driving progress and innovation in various fields. The creation of new distributions and the enhancement of existing ones play a crucial role in optimizing analyses and continually improving the accuracy of statistical inference processes. In this context, an alternative in statistical modeling is the use of the Inverted Weibull distribution, which is widely used in reliability studies and survival analysis. One way to enhance and extend these distributions is the use of distribution generators. The Cosine Type I Generator (TIC) is a type of generator that does not introduce complexity into the resulting parametric space, as it does not add parameters to the new distribution while still capturing different modeling situations. Therefore, the aim of this work is to propose a new distribution based on the Inverted Weibull with the use of the TIC generator, referred in this work as Cosine Type I Inverted Weibull with parameters α and λ and denoted as TICIW. A study on the parameter estimation of this new distribution was conducted using the maximum likelihood estimation method and evaluated through Monte Carlo Simulation with 10000 replications and sample sizes . To assess the Maximum Likelihood Estimator (MLE), performance measures such as mean, bias, relative bias (RB), standard deviation (SD), mean squared error (MSE), skewness coefficient (SC), and kurtosis (K) were used. The result of this simulation study, supported by the measures used, confirmed the accuracy and precision of the estimators derived from the new TICIW distribution.
- Cleber Bisognin
- Augusto Maciel da Silva