Research on Attribute Discretization for One Airborne Radar Intelligence System Based on Improved FCM Clustering

摘要

To address the information loss problem in the process of discretizing continuous attributes for one airborne radar intelligence system, this paper proposes a discretization method based on the improved FCM(Fuzzy c-means) algorithm through an analysis of the existing attribute discretization methods. This method improves the original algorithm by introducing the fuzzy decision theory and the validity index which is based on the geometric structure of the dataset, and uses this method to make the continuous attributes fuzzy. It overcomes the shortcomings that the conventional fuzzy clustering method must use two prespecified parametersȁ4;the weighted index m and the number of clusters c, and does not consider the specified attribute values of the dataset. We use the proposed method to discretize the continuous attributes in the airborn radar intelligence database and experimental results show that the scheme is feasible and effective.