Structural Time-Series Models for Forecasting Yield of Promising Varieties of Wheat Crop in Chhattisgarh

A univariate structural time series model based on the traditional decomposition into trend, seasonal and irregular components is measured. Purpose of present paper is to discuss Structural Time Series Model (STM) methodology utilized for modelling time-series data in the present of trend, seasonal and cyclic fluctuations. Structural time series model are formulated in such a way that their components are stochastic, i.e. they are regard as being driven by random disturbances. The study mainly confined to secondary collected for a period 2009–10 to 2014–15 data of promising varieties of Wheat yield. As these techniques, it may be mentioned that models are fitted to the data and coefficient parameter value obtained on the basisof the model are compared with the actual observation for assessing the accuracy of the fitted model. To validate the forecasting ability of the fitted models, for the three years with upper andlower limit. The Lok-1 variety have minimum AIC (18.84) and BIC (16.13) obtained among 9 high yielding varieties followed by HW-2001 with AIC (20.23) and BIC (17.52) respectively. The maximum AIC (27.09) and BIC (24.39) obtained for the variety of HI-1544.