Representative Selection by Message Passing Graph Clustering

초록

We describe a new method of voting system in social networks. We suggest a sequence of continuous support via a social network after electing representatives in the network that is different from the typical majority voting. In other words, this paper suggests the method of elected representatives using network clustering but just counts voting. On the network structure, sending messages reflected real value, we address to cluster for representatives by whom the representatives support in the graph. In the experiment our algorithm outperformed conventional approaches in social network synthetic dataset as well as real dataset.