Abstrak |
: |
This paper discusses about developing generalized linear model (GLM) of Gamma-Pareto (G-P) distribution. It is shown that G-P is a member of exponential family distribution. Consequently, we can develop GLM based on G-P. With as response variable, the GLM has inverse, identity, and log link functions. Dispersion parameter, deviance, and Akaike information criteria (AIC) of the model are derived in this paper. We also provide the iterative weighted least squares (IWLS) for estimating the model’s parameters. |