| 主权项 |
1. A method comprising:
identifying, by a system having a processor, concepts in a plurality of clusters of documents; identifying, by the system for a first cluster of the plurality of clusters, groups of connected concepts of the first cluster; computing, by the system, interestingness measures for each of the groups of connected concepts in the first cluster, wherein computing the interestingness measures for each group of connected concepts comprises computing different types of interestingness measures, the different types of interestingness measures for a first group of connected concepts of the groups of connected concepts comprising a first type of interestingness measure representing a likelihood of a connection between the concepts of the first group of connected concepts, and a second type of interestingness measure representing an importance of the concepts of the first group of connected concepts; deriving, by the system, an interestingness measure for the first cluster based on the interestingness measures for the corresponding groups of connected concepts; and re-iterating the identifying of the groups of connected concepts, the computing, and the deriving for other clusters of the plurality of clusters, to produce respective interestingness measures for the other clusters. |
