Package edu.umass.cs.mallet.projects.seg_plus_coref.condclust.pipe

Class Summary
AllLinks Subsumes ClosestSingleLink, AverageLink, FarthestLink to save computation
AverageLink Feature is similarity between node and closest node in cluster, as determined by the classifier
ClosestSingleLink Feature is similarity between node and closest node in cluster, as determined by the classifier
ClusterHomogeneity Feature is average within-class similarity.
ClusterSize Feature is size of cluster...to penalize large clusters.
FarthestSingleLink Feature is similarity between node and farthest node in cluster, as determined by the classifier
ForAll Sets a feature for each element of "fields" that is true if it is an exact string match for Node and for all Nodes in the Cluster
NNegativeNodes Feature is true if there exist at least N nodes in Cluster that have a negative edge weight with Node
NodeClusterPair2FeatureVector  
PaperClusterPrediction Feature is the output of the paperClusterClassifier
ThereExists Sets a feature for each element of "fields" that is true if it is an exact string match for Node and for some Node in the Cluster
ThereExistsMatch More specific "does there exist a node such that" questions
VenueClusterPrediction Feature is the output of the paperClusterClassifier
VenuePaperCluster2FeatureVector