Gaussian and Non-Gaussian Autoregressive Time Series Models with Rainfall Data
Sukhpal Kaur1, Madhuchanda Rakshit2
1Sukhpal Kaur, Research scholar, Department of Computer Science, Guru Kashi University Punjab, India.
2Dr. Madhuchanda Rakshit, Assistant Professor, Department of Computer Science, Guru Kashi University Punjab, India.
Manuscript received on September 23, 2019. | Revised Manuscript received on October 05, 2019. | Manuscript published on October 30, 2019. | PP: 6699-6704 | Volume-9 Issue-1, October 2019 | Retrieval Number: A1994109119/2019©BEIESP | DOI: 10.35940/ijeat.A1994.109119
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: The Gaussian and non- Gaussian autoregressive models are used in this paper for analyzing time series data. The autoregressive time series models with various distributions are considered here for analyzing the annual rainfall of Punjab, India. Three different types of autoregressive models are applied here for analyzing data namely autoregressive model with Gaussian, Gamma and Laplace distribution. For the goodness of fit the chi – square test is applied and the best fitted distribution is obtained for the data. Next the stationarity of data is checked, after that models are applied on data for comparing three distributions of AR models and lastly the best fitted model is obtained. The residual checking of selected model is also discussed and forecast the best fitted model based on simulated response comparison.
Keywords: Chi-square test, Laplace AR model, Gamma AR model, Gaussian AR model, NRMSE.