Application of nonparametric additive model for input-response analysis in arecanut

Abstract

The relationships between the inputs and response are very complex in crop production models. Problems like nonlinear relationships between inputs and responses, nonexistence of proper functional form to represent relationship between inputs and response variable and multi-collinearity are very common in crop production data. The traditional multiple linear regression technique may not be adequate in many situations to explain input output relationship. In this paper we have used nonparametric additive regression model to explain input output relationship in arecanut. The comparative study shows that the nonparametric additive modeling technique performed much better than the multiple linear regression technique to explain the input response relationship. The estimated values of the component functions provide the mean response of input variables on the yield of arecanut. The optimum value of the input variables is obtained from the graphical representation of the component functions. The present analysis of data from the two districts show that input response relationship vary depending on the agro-climatic conditions of the locations.