Pantnagar Journal of Research

Abstract

The present study was carried out to bring the realistic nature of modeling variables, using non-linear regression approach on area and production of sugarcane in Uttar Pradesh, using the secondary data of area and production from1980 to 2017. Six different methods of nonlinear regression models as Logistic, Power, Sinusoidal, Richards, Rational and Ratkowsky were used. Levenberg-Marquardt technique was used to estimate the unknown parameters of the nonlinear regression models. To select a best fitted model by using , RMSE, MAE, and residual assumption tests such as Run test, Shapiro-Wilks test were carried out. For area, production and productivity it was found that Rational model had the highest values i.e. (85 per cent, 83 per cent, 75 per cent) and lowest RMSE (84.4, 9992.7, 3546.4). Hence, Rational model is the most suitable among the fitted nonlinear model which can be used for further trend analysis of sugarcane in Uttar Pradesh.