Clustering algorithms for unsigned networks which have only positive edges have been studied intensively. However, many networks arise in situations where relationships are best expressed by both positive and negative values and in such situations unsigned networks with only positive edges are inadequate. Currently, there are few clustering algorithms for signed networks due to the fact that such algorithms not only require the clusters of a signed network to be dense (as required in the unsigned case), but also that there should be as many negative edges as possible between clusters. In this talk we will discuss signed network clustering and propose a clustering algorithm, EB+D, for signed networks, where both the betweenness of edges and the density of subgraphs are considered. A hierarchically nested system is constructed to illustrate the inclusion relationships of clusters. We will describe some results using several classical social networks and hundreds of synthetic data sets