Seminario - Reparameterized slashed lognormal regression model: Diagnostics and application to mineral data

 Jeremias Leão

Resumo: In this article, we propose a slashed lognormal regression model based on a reparameterization defined in terms of the mean, precision, and shape parameters. The model is suitable for situations in which the response variable is continuous, positively skewed, exhibits a flexible shape, and presents high kurtosis and heteroskedasticity. The proposed methodology includes parameter estimation, hypothesis testing for the precision parameter, residual analysis, and influence diagnostics. A Monte Carlo simulation study is carried out to evaluate the performance of the estimators and residuals in finite samples. Finally, the applicability of the model is illustrated using a mineral concentration dataset.

Link:  https://teams.microsoft.com/meet/255441543305343?p=reYtulHWSQS4X7ykJ5