Time Series Analysis of General Marine Fisheries

Fish production forecasting is a crucial factor in fishery resource management and development of fisheries. Fish production depends on very various effects and issues so that it is complicated to predict it considering all relevant factors. It is ARIMA model that can fairly overcome these limitations because ARIMA model can perform forecasting despite insufficient information on the relevant factors. However, ARIMA model could make it difficult to discriminate the parameter. In order to complement the defects, we developed a looping simulation program that could generate the results of all fisheries cooperatives by setting the range of the parameter. In addition, ARIMA model generally shapes a single model for a single target variable. But the model could affect seriously forecasting result when it does not properly infer coefficients of the model. To overcome the defect, We proposed a forecasting prospect through double time series models that mix ARIMA model and SARIMA(seasonal ARIMA) model. Consequently, We presented fish production forecasting model of anchovy, mackerel, hairtail and yellow corbina.