The Establishment of Northwest Pacific Ommastrephes bartramii Fishing Ground Forecasting Model Based on Naive Bayes Method
Northwest Pacific is an important fishing ground of Ommastrephesbartramii for China.The accurate prediction of fishing grounds is critical to energy saving and catch increasing.In this study,a squid fishing ground forecasting model was built with Naive Bayesian method according to the historical catching,spatio-temporal and marine environmental data of this region.The environmental data included sea surface temperature (SST),concentration of chlorophyll ((CHL),SST gradient strength and CHL gradient strength collected from 2002 to 2011.In order to satisfy the conditional independence assumption of naive Bayes method,the independent component analysis was carried out.By maximizing the Cohen's Kappa coefficient,three-type CPUEs prior probabilities were determined for the Naive Bayesian forecasting model.As a test,the forecasted fishing ground maps were overlapped with the practical fishing grounds found from July to November 2012.Such an overlapping showed that the average comprehensive accuracy of forecasted fishing grounds was 69.9 ％,indicating that the Naive Bayesian model was effective for and feasible in forecasting the squid fishing grounds in northwest Pacific.
国家科技期刊平台
NSTL
万方数据
维普
