Search

Graph clustering acts as a critical topic for solving decision situations in networks. Different node clustering methods for undirected and directed graphs have been proposed in the literature, but less attention has been paid to the case of attributed weighted multi-edge digraphs (AWMEDiG). Nowadays, multi-source and multi-attributed data are used increasingly in decision sciences; however, traditional methods usually consider single-attributed and single-view data as the input. This type of directed network, whose nodes are described by a list of attributes and directed links are viewed as directed multi-edge, is a new challenge to graph clustering. This paper proposes a new approach to detecting and evaluating clusters of AWMEDiG based on the maximum clique method. Our algorithm first converts the given AWMEDiG into a new weighted graph. This transformation is carried out through a new structural-attributed similarity measurement, an improved version of our previous model. Then, the concept of the maximum clique is adopted to complete the clustering task. The main objective of this new approach is to improve the clustering quality by grouping the highly cohesive and homogeneous vertices. So, the entropy strongly tends to be zero. Moreover, since the number of clusters is not predefined, it has the ability to find out the natural number of clusters within a graph. The performance of the proposed algorithm is tested on a synthetic, provided for a three-view attributed weighted directed network including 50 nodes, and two real-world datasets: UK faculty and Seventh graders. The clustering results are analyzed based on the clusters' entropy and density, meanwhile the precision and F-score indexes measure the accuracy of them. [ABSTRACT FROM AUTHOR]

This paper treats with a decision tree consisting of RBF neural networks in the nodes to decrease the classification time and to improve accuracy as well as generalization. We propose two knowledge transferring mechanisms between nodes to reduce the duplicate computations in training process. The resulted classifier is titled as ensemble of RBF neural networks in decision tree structure with knowledge transferring or shortly ERDK. We accelerate ERDK by applying a cut-point mechanism to prune its tree structure. The results on a great number of benchmark datasets show that ERDK provides promising results. Particularly, AUCarea, F-measure, and G-mean for ERDK are 95, 90, and 90%, respectively. As a case study, we finally apply ERDK on Britain incident dataset. [ABSTRACT FROM AUTHOR]