MODELING OF INDIAN OCEAN DIPOLE INDEX VIA W-GARCH METHODS

To pursue a superior modeling tools for improving the predictive efficacy in Indian Ocean Dipole Index (IOD) using a hybrid time series model. A methods which makes use of wavelets together with a time series model, GARCH, is fitted in order to gain the benefit of a hybrid model in Ocean atmospheric interaction variability data on tropical Indian Ocean. Initially classical time series analysis is performed. Following that, volatility feature is used in Generalized Autoregressive Conditional Heteroscedasticity (GARCH) accompanied with wavelet smoothing tools that has been experimented on the monthly data, collected from January 1870 to December 2018 of dipole mode index on tropical Indian Ocean. Results: W-GARCH(1,1) model about student-t distribution has been found as the most suitable model with the least AIC, - 28.10682. The model was validated by the training and test data set by Mc-LeodLi test, QQ plot and many others. Results are showed that the proposed methods, WGARCH(1,1), has made efficient results compared with the GARCH. Therefore, modeling and forecasts based on the methods could assist in improving the climate forecast system in the fields that are influenced by IOD