Difference between revisions of "Link Prediction"
From Cohen Courses
Jump to navigationJump to search (Created page with 'Given a snap shot of current network over a set of nodes, link prediction is method of identifying which new links that are likely to occur in the near future. In a graph with hu…') |
|||
| (One intermediate revision by the same user not shown) | |||
| Line 1: | Line 1: | ||
| − | Given a snap shot of current network over a set of nodes, link prediction is method of identifying which new links that are likely to occur in the near future. In a graph with huge number of nodes, identifying the potential nodes that a given node gets attached to in near future is a non trivial task. Typically, these problems are solved by coming up with some kind of "proximity" measure (like Page-Rank score) for the nodes in the graph, and predicts that the links being formed with nodes having higher proximity measures. | + | Given a snap shot of current network over a set of nodes, link prediction is method of identifying which new links that are likely to occur in the near future. In a graph with huge number of nodes, identifying the potential nodes that a given node gets attached to in near future is a non trivial task. Typically, these problems are solved by coming up with some kind of "proximity" measure (like Page-Rank score) for the nodes in the graph, and predicts that the links being formed with nodes having higher proximity measures. Seminal paper which describes the problem is here [http://www.cs.cornell.edu/home/kleinber/link-pred.pdf Kleinberg] |
| + | == Relevant Papers == | ||
| + | |||
{{#ask: [[AddressesProblem::Link Prediction]] | {{#ask: [[AddressesProblem::Link Prediction]] | ||
| ?UsesMethod | | ?UsesMethod | ||
| ?UsesDataset | | ?UsesDataset | ||
}} | }} | ||
Latest revision as of 17:24, 1 April 2011
Given a snap shot of current network over a set of nodes, link prediction is method of identifying which new links that are likely to occur in the near future. In a graph with huge number of nodes, identifying the potential nodes that a given node gets attached to in near future is a non trivial task. Typically, these problems are solved by coming up with some kind of "proximity" measure (like Page-Rank score) for the nodes in the graph, and predicts that the links being formed with nodes having higher proximity measures. Seminal paper which describes the problem is here Kleinberg
