Using Autoregressive Integrated Moving Average (ARIMA) Technique to Forecast the Production of Kharif Cereals in Odisha (India)

Dash, Abhiram; Mangaraju, A.; Mishra, Pradeep Kumar; Nayak, Harapriya

doi:10.9734/cjast/2020/v39i930619

Cited by 3 publications

(1 citation statement)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The future forecast values of production for kharif cereals is found to decrease which might be due to the decrease in forecast values for area. The forecast values of area and yield of kharif cereals obtained by fitting ARIMA model also shows the similar result but for a shorter period from 2016-17 to 2018-19 [9].…”

Section: Resultssupporting

confidence: 68%

Forecasting of Kharif Cereal Production in Odisha by Using Spline Regression Technique

Basanaik

Dash

2021

IJPSS

Self Cite

View full text Add to dashboard Cite

Cereals are prime determinant of agricultural status of the state mainly during kharif season. Forecasting of the production of kharif cereals is of utmost importance to formulate the agricultural policy and strategy of the state. The ARIMA model can be reliably used to forecast for short future periods because uncertainty in prediction increases when done for longer future periods. The predictions obtained from the ordinary regression model are valid only when the relationship between the independent variables and the dependent variable does not change significantly in the future period which can be rarely assumed. It is expected that the spline regression will overcome the respective discrepancies in both ARIMA and ordinary regression techniques of forecasting with the assumption that the future period which needs forecasting follows the same pattern as the last partitioned period. The entire period of data is split into different periods based on the scatter plot of the data The suitable regression models, such as, linear, compound, logarithmic and power model are fitted to the data on area and yield of kharif cereals by using the training set data. Selection of best fit model is done on the basis of overall significance of the model, model diagnostic test for error assumptions and model fit statistics. The selected best fit model is then cross validated with the testing set data. After successful cross validation of the selected best fit models, they are used for forecasting of the future values for their respective variables. The models found to be best fit and thus selected for cross validation purpose are compound spline model for both area and yield of kharif cereals respectively. Forecasting of area, yield and hence production of kharif cereals for six years ahead i.e., for the year 2020-21 to 2025-26 by using the selected best fit model after successful cross validation. The forecast values for production of kharif cereals are found to decrease despite increase in forecast values of yield which is due to decrease in forecast value of area.

show abstract

Section: Resultssupporting

confidence: 68%