Abstract
Although Akaike information criterion is the most popular selection criterion in the literature, it gives inconsistent results in determining the correct model according to other selection criteria in transmuted distribution families. We motivate a new extension of the Akaike information criterion in the solution of this problem. In this paper, we suggest a new selection criterion as an alternative to the Akaike information criterion for the family of transmuted distribution. We discuss special cases of this family based on exponential distribution. A Monte Carlo simulation study is considered to compare the performances of this new criterion with the Akaike information criterion. Also, a numerical example is presented.
Keywords
Akaike information criterion
Bayesian information criterion
Mathematics
Exponential family
Selection (genetic algorithm)
Exponential distribution
Model selection
Statistics
Applied mathematics
Exponential function
Mathematical optimization
Computer science
Artificial intelligence
Mathematical analysis
Subject Areas
Statistical Distribution Estimation and Applications
·Statistics and Probability ·Physical Sciences
Probability and Risk Models
·Management Science and Operations Research ·Social Sciences
Probabilistic and Robust Engineering Design
·Statistics, Probability and Uncertainty ·Social Sciences
Loading…