Abstract: A system and method for redistributing interest in a hierarchical data structure representation is provided. A data representation of a hierarchical data structure includes depth levels having one or more blocks of sibling nodes associated with node interest characteristics. Two or more of the sibling nodes are combined as aggregates into elided subsets. One of the elided subsets is selected and expanded to reveal the aggregated sibling nodes. A breadth of the depth level in which the block is located is determined. A further determination that the breadth exceeds a constrained breadth extent for the data representation is made. The breadth is decreased by ordering the sibling nodes in the depth level by their node interest characteristic and by combining at least a portion of the sibling nodes into the elided subset based on the respective node interest characteristics such that the breadth does not exceed the constrained breadth extent.

Abstract: A system and method for redistributing interest in a hierarchical data structure representation is provided. A data representation of a hierarchical data structure includes depth levels having one or more blocks of sibling nodes associated with node interest characteristics. Two or more of the sibling nodes are combined as aggregates into elided subsets. One of the elided subsets is selected and expanded to reveal the aggregated sibling nodes. A breadth of the depth level in which the block is located is determined. A further determination that the breadth exceeds a constrained breadth extent for the data representation is made. The breadth is decreased by ordering the sibling nodes in the depth level by their node interest characteristic and by combining at least a portion of the sibling nodes into the elided subset based on the respective node interest characteristics such that the breadth does not exceed the constrained breadth extent.

Abstract: Techniques for estimating user interest in graph structures are provided. A graph structure containing at least two nodes, a threshold disinterest value and at least one interesting node within the graph structure are determined. Each determined interesting node is added to a set of active nodes. Adjacent nodes connected to the set of active nodes and associated with Degree-Of-Interest values more interesting than the threshold disinterest value are in turn added to the set of active nodes until no additional adjacent connected nodes have a Degree-Of-Interest value more interesting than the threshold value. A new visualization of the graph structure is determined based on the nodes in the set of active nodes. The interesting nodes may be determined based on specific indications of interest in a node, such as a mouse selections, or may be based on the user's focus of attention within the graph based information structure.

Abstract: Techniques for clustering user sessions using multi-modal information including proximal cue information are provided. The topology, content and usage of a document collection or web site are determined. User paths are then identified using longest repeating subsequence techniques. An information need feature vector is determined for each significant user path. Further, other feature vectors and proximal cue vectors for each document or web page in the significant path are determined. The other feature vectors include a content feature vector, a uniform resource locator feature vector, an inlink feature vector and an outlink feature vector, among others. The feature vectors and the proximal cue vectors are combined into a multi-modal vector that represents a user profile for each significant user path. The multi-modal vectors are clustered using a type of multi-modal clustering such as K-Means or Wavefront clustering.

Abstract: Techniques for clustering user sessions using multi-modal information including proximal cue information are provided. The topology, content and usage of a document collection or web site are determined. User paths are then identified using longest repeating subsequence techniques. An information need feature vector is determined for each significant user path. Further, other feature vectors and proximal cue vectors for each document or web page in the significant path are determined. The other feature vectors include a content feature vector, a uniform resource locator feature vector, an inlink feature vector and an outlink feature vector, among others. The feature vectors and the proximal cue vectors are combined into a multi-modal vector that represents a user profile for each significant user path. The multi-modal vectors are clustered using a type of multi-modal clustering such as K-Means or Wavefront clustering.