A log-location-scale lifetime distribution with censored data: Regression modeling,residuals, and global influence |
| |
Authors: | Francisco Louzada Vitor A. A. Marchi |
| |
Affiliation: | 1. Department of Statistics, Universidade Federal de S?o Carlos, S?o Carlos, Brazillouzada@icmc.usp.br;3. Department of Statistics, Universidade Federal de S?o Carlos, S?o Carlos, Brazil |
| |
Abstract: | In survival analysis applications, the presence of failure rate functions with non monotone shapes is common. Therefore, models that can accommodate such different shapes are needed. In this article, we present a location regression model based on the complementary exponentiated exponential geometric distribution as an alternative to the usual bathtub, increasing, and decreasing failure rates in lifetime data. Assuming censored data, we consider the maximum likelihood inference for analysis, graphical verification for residuals, and test statistics for influential points. |
| |
Keywords: | Censored data complementary risks exponential distribution failure rate function survival analysis. |
|
|