A new feature selection method in fishery information processing

Fishery information processing can help fishery researchers obtain the needed information easily and quickly. The current information processing techniques have not solved the problem of high dimensional features in fishery information processing. In this paper, a feature selection method for fishery texts based on SVM-RFE was put forward in view of the characteristics of fishery texts. It removed the redundant information in text feature space and reduced the feature dimensions effectively. Three corpora were employed to verify the proposed method and the comparison with the traditional feature selection method was performed. The experimental results show that the method proposed in this paper can improve precision rate and recall rate of fishery information processing with the lower dimensional features, providing an effective way for fishery information processing.