Weather prediction, Sensex prediction, sales prediction, and so on, are some of the common problems that are of interest to most who look up to Statistics or Machine Learning methods. The purpose is, of course, to get the prediction for the next few periods using models with reasonable accuracy. Weather prediction helps in planning for the vocational trip, Sensex prediction helps in investment planning, and sales prediction in an optimum inventory planning. The common structure among each of the three problems is that the observations are available, in general, at equally spaced time/epoch. The observations may have been obtained on a daily, weekly, or monthly basis and we will refer to such data as time series data. The observations are collected over a long time in the past and we believe that we have captured enough characteristics/traits of the series so that analytical models built on such historic data will have predictable power and we can obtain fairly accurate forecasts...