Rasaki Olawale Olanrewaju1, Johnson Funminiyi Ojo2, Adekola Lanrewaju Olumide3
1. Department of Mathematical Sciences, Pan African University Institute for Basic Sciences, Technology and Innovation. P.O. Box 62000-00200, Nairobi, Kenya.
2. Department of Statistics, University of Ibadan, Ibadan, P.O. Box 200284, Oyo state, Nigeria.
3. Department of Physical Sciences, Bells University of Technology, Ota, Nigeria.
Abstract: We study the assortment of autoregressive random processes via a transmuted Gamma distributed noise. We consider a transmuted re-parameterization of the Gamma parameters in terms of μ and σ2, afterwards ascertained that the transmuted Gamma is a proper probability density function, then proceeded to spelt-out the structural form and traits of the Gamma Mixture Autoregressive generalization in its k-components. The mean and variance of the Gamma Autoregressive model were ascertained coupled with its first and second-order stationarity. The ingrained k-components’ autoregressive coefficients, re-parameterization Gamma coefficients, k-regime transitional weights were estimated via Expectation-Maximization (EM) algorithm. However, some step ahead predictions were derived as well as the model sub-setting estimation via Levinson-Durbin recursive technique.
Keywords: Expectation-Maximization (EM) algorithm, Gamma Autoregressive model, k-components, k-regime transitional weights, Levinson-Durbin recursive, Transmuted.
Pages: 183 – 202 | Full PDF Paper