Statistical modelling for describing fish production in India

Abstract

Various statistical modelling approaches. viz. polynomial function fitting approach. non-linear mechanistic growth modelling approach, and ARIMA time-series analysis approaches are discussed for describing marine, inland as well as total fish production of the country during 1950-51 to 1994-95. The underlying assumptions are examined by statistically sound procedures. In view oftht; presence oftht; autocorrelations among observations in the 3 data sets, it is found that the ARIMA time-series analysis approach is most appropriate. The identified models are then used to forecast the future fish production.