Forecasting of fish production from ponds – a nonlinear model approach

Abstract

Growth of fish is affected by many factors which are usually time dependent. The present study attempts to develop a forecasting model of fish production based on growth models. The concept of partial reparameterisation by expected-value parameters to mitigate nonlinear behaviour of parameters is highlighted. A reparameterised version of the Gompertz model is proposed for forecasting fish production, which is illustrated with the average growth data of grass carp rearing in polyponds of upland region. Two months ahead forecasting of grass carp production in the 11th and 12th months of rearing period is best predicted by the proposed model.