Nonlinear regression: a realistic modelling approach in horticultural crops research
2003
Venugopalan, R. | Shamasundaran, K.S. (Indian Institute of Horticultural Research, Bangalore (India))
Nonlinear statistical models play a very important role in almost all branches of agricultural/horticultural and biological sciences due to the existence of wide population fluctuations and complex non-linear inter-relationship among variables of interest. This article brings out the realistic nature of modelling such variables using non-linear regression approach. Four different methods of non linear regression are discussed and measures of goodness of fit are presented with a view to develop a suitable non-linear model for describing data pertaining to the period 1960-61 to 1976-77, on average fruit yield of coorg mandarin trees. Using the Gompertz model selected, it was inferred that 94 percent of tha carrying capacity (maximum sustainable yield) had already been achieved by the year 1977 and hence there is little scope for its survival in Coorg region.
显示更多 [+] 显示较少 [-]