Present a link prediction algorithm in temporal uncertain social networks.

We transform the link prediction in uncertain network to a random walk in a deterministic one.

We compute the similarity scores between a node and its neighbors within a subgraph, instead of computations in the whole network.

Present an O(|V|³) method to compute the similarity matrix, instead of enumerating the 2^|E| possible worlds in the network.

We propose a method for integrating temporal and global topological information in temporal uncertain networks.