A statistical modeling approach for forecasting of pulses production in Tamil Nadu

Abstract

The present study was conducted to check the trend analysis of area, production and production for pulse crops in select districts of Tamil Nadu. Time series data on area, production and productivity of black gram and green gram in select districts of Tamil Nadu were from 1997-98 to 2015-2016. The objective of this study were to fit different trend equations like linear, non-linear and fuzzy time series models for pulse crops like Black gram and Green gram for select districts of Tamil Nadu and then select the most appropriate model. The best-fitted model for future projection was chosen based upon highest Theil’s U-Statistic, coefficient of determination (R2) and with least MAPE, MAE and RMSE values. The findings of the study are based on the least RMSE, MSE and MAPE value, among all the linear and nonlinear models for black gram production quadratic model was best fitted model for Cuddalore district and the forecasted value during 2020 AD would be 101.60 thousand tonnes. For Nagapattinam and Trivarur districts, Cubic model was best fitted model. The forecasted value for 2020 AD would be 38.99 thousand tonnes (Nagapattinam district) and 131.79 thousand tonnes (Trivarur district). For green gram production Exponential model was the best fitted model for Tuticorin district and Cubic model was appropriate for Trivarur and Nagapattinam districts. The forecasted value for green gram production using Exponential model for the year 2020 AD would be 35.06 thousand tonnes for Tuticorin district. The forecasted value for green gram production using cubic model for the year 2020 AD would be 109.93 thousand tones for Trivarur district and 59.27 thousand tonnes for Nagapattinam district.