Abstract: A method assigns a score to each node in a directed graph. Nodes in the graph represent autonomous entities, and links denote opinions entities hold of each other. Scores are assigned based on either a deterministic iterative method or a random walk. Both methods are able to take negative opinions into account by assigning negative reputation to a node in proportion to the positive reputation of the node that links to it with a negative opinion, and also assigning a separate kind of negative reputation to nodes that have a positive opinion of a node with either kind of negative reputation. The random walk method also solves the “rank sink” problem of previous methods by disallowing any single traversal from visiting any particular node more than once within a certain interval.

Abstract: A method assigns a score to each node in a directed graph. Nodes in the graph represent autonomous entities, and links denote opinions entities hold of each other. Scores are assigned based on either a deterministic iterative method or a random walk. Both methods are able to take negative opinions into account by assigning negative reputation to a node in proportion to the positive reputation of the node that links to it with a negative opinion, and also assigning a separate kind of negative reputation to nodes that have a positive opinion of a node with either kind of negative reputation. The random walk method also solves the “rank sink” problem of previous methods by disallowing any single traversal from visiting any particular node more than once within a certain interval.