Knowledge Discovery and Representation for Fishery Forecasting

In the marine industry there has always been immense research interest in maximizing accuracy of fishery forecasting. The fishery knowledge have a great impact on the accuracy, so this paper proposes a new knowledge discovery and representation model for obtaining and representing fishery knowledge, which takes a 3 step process. Firstly, it extracts static knowledge from database by SVM classifier and fuzzy classifier. Secondly, it uses extension data mining method to transfer static knowledge into dynamic knowledge. Thirdly, it establishes an ontology knowledge base by utilizing a mapping mechanism between the dynamic knowledge and ontology. Using the proposed model building procedure, the authors implemented a prototype system for fishery forecasting. Experimental results show that the proposed method is effective and efficient.