STATISTICAL MODEL FOR FORECASTING OF AREA AND PRODUCTION OF GARDEN PEA AND CAULIFLOWER IN HIMACHAL PRADESH
Loading...
Date
2020-08
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
UHF,NAUNI
Abstract
ABSTRACT
The present investigation entitled “Statistical model for forecasting of area and production of
Garden Pea and Cauliflower in Himachal Pradesh” was carried out on the time series data from 1996-
2016 (21 years including 2016). Time series data was collected from Directorate of Agriculture, Shimla.
Regression analysis was carried out using linear and non linear model. Also autoregressive models were
fitted based on the significance of autocorrelation coefficient. Adjusted-R square ( ), root mean square
error (RMSE), Theil’s U statistic and F-chow statistic were used for selection of model. Quadratic model
was found best fit for the estimation of area and production of garden pea. Cubic model was found best fit
for prediction of area and production of cauliflower. ARIMA models were fitted for the time series data.
Series was stationarized by double differencing of the data. Significant spikes in ACF and PACF plots
were used to identify the number of moving average and autoregressive terms in the time series data.
Akaike’s information criterion (AIC), RMSE, mean absolute error (MAE), mean absolute percent error
(MAPE) and Ljung-Box coefficient were used to select the best model. ARIMA (2, 2, 0) was found best
fit for estimation of both area and production of garden pea. ARIMA (1, 2, 1) and ARIMA (0, 2, 1) were
found best fit for estimation of area and production of cauliflower respectively.