基于集体影响和改进的重组方法识别重要节点.ppt

Identifying vital nodes based on Collective Influence and Improved Reinsertion Solution for Datacastle Master Competition TeamID: zhfkt Fengkuangtian Zhu Datacastle Master Competition Task The target of Competition is to give a ranking list of nodes according to their importance. The top-ranked nodes will have more importance. We can remove the nodes from the top-ranked ones in the ranking list and calculate the size of giant component after each removal, which will break down the network into many disconnected pieces. Therefore, the better algorithm, the sooner the network will collapse to the zero giant component with smaller count of provided nodes. Collective Influence Morone F, Makse H A. Influence maximization in complex networks through optimal percolation[J]. Nature, 2015, 524(7563): 65-U122. Morone F, Min B, Bo L, et al. Collective Influence Algorithm to find influencers via optimal percolation in massively large social media[J]. Scientific reports, 2016, 6. Collective Influence Collective Influence (CI) algorithm using optimal percolation for localizing the minimal number of influential nodes. The problem of finding the minimal set of influencers can be mapped to the optimal percolation. CI will calculate the value of each node in the following Formula and remove the highest value of nodes. Collective Influence Collective Influence A) Removing the nodes with the largest collective influence (the influencers) allows for efficient dismantling of networks. B) removing 6 nodes with the highest degree leaves the network mostly intact C) removing 4 top influencers leaves subgraphs that are smaller. Reinsertion Reinsertion adds back one of the removed nodes, which is chosen such that, if once rein- serted, it joins the smallest number of clusters. When the node is reinserted reinsertion also restores the edges with its neighbors which are in the network (but not the ones with neighbors not yet reinserted, if any). The procedure is repeated until all the nodes