No abstract

Several previous works proposed techniques to detect outliers in graph data. Usually, some complex dataset is modeled as a graph and a technique for detecting outliers in graphs is applied. The impact of the graph model on the outlier detection capabilities of any method has been ignored. Here we assess the impact of the graph model on the outlier detection performance and the gains that may be achieved by using multiple graph models and combining the results obtained by these models. We show that assessing the similarity between graphs may be a guidance to determine effective combinations, as less similar graphs are complementary with respect to outlier information they provide and lead to better outlier detection.

Dense subgraphs detection is a well known problem in Computer Science. Hierarchical organization of graphs as dense subgraphs, however, goes beyond simple clustering as it allows the analysis of the network at different scales. Despite the fact there are several works on hierarchical decomposition for unipartite graphs, only a few works for the bipartite case have been proposed. In this work we explore the problem of hierarchical decomposition of bipartite graphs. We propose an algorithm which we call weighted linking that produces denser and more compact hierarchies. The proposed algorithm is evaluated experimentally using several datasets and provided gains on most of them.

No abstract

Dense subgraph detection is a well-known problem in graph theory. The hierarchical organization of graphs as dense subgraphs, however, goes beyond simple clustering, as it allows the analysis of the network at different scales.Although there are several hierarchical decomposition methods for unipartite graphs, only a few approaches for the bipartite case have been proposed. In this work, we explore the problem of hierarchical decomposition for bipartite graphs.We propose an algorithm called Weighted Linking that identifies denser and more compact hierarchies than the state of the art approach. We also propose a new score to help choose the best between two hierarchical decompositions of the same graph.The proposed algorithm was evaluated experimentally using six real-world datasets and identified smaller and denser hierarchies on most of them.