Computer-implemented system and method for generating clusters for placement into a display
View Patent ↗A computer-implemented system and method for generating clusters for placement into a display is provided. A set of clusters is generated from a document set. A single cluster of related documents from the document set is obtained and at least one new cluster is added. One such document in the set is compared to the cluster. A difference in distance between the document and a common origin and the cluster and the common origin is determined. The document is designated as the new cluster when the difference fails to satisfy a predetermined threshold. One or more cluster spines each having two or more clusters placed along a vector are placed into a display. The clusters along each spine are identified as similar and the clusters of one such spine are also similar to further clusters located along a further spine having a small cosine rotation from that cluster spine.
1. A computer-implemented system for generating clusters for placement into a display, comprising:
a database to store a document set; and
a server comprising memory, a central processing unit, an input port to receive the document set from the database, and an output port, wherein the central processing unit is configured to:
generate a set of clusters from the document set by obtaining a single cluster of related documents from the document set and adding at least one new cluster comprising comparing one such document in the set to the cluster, determining a difference in distance between the document and a common origin and the cluster and the common origin, and designating the document as the new cluster when the difference fails to satisfy a predetermined similarity threshold;
place into a display one or more cluster spines each comprising two or more of the clusters in the set placed along a vector; and
identify the clusters along each spine as similar, wherein the clusters in one such spine are also similar to further clusters located along a further spine having a small cosine rotation from that cluster spine.
2. A system according to claim 1 , wherein the display is three-dimensional and each cluster represents multi-dimensional data modeled in the three-dimensional display.
3. A system according to claim 1 , further comprising:
an identification module to identify for each cluster in the set a center and radius.
4. A system according to claim 3 , wherein the center of each cluster is located at a fixed distance from the common origin.
5. A system according to claim 3 , wherein the radius of each cluster in the set is independent of the other clusters and reflects a relative number of the related documents in that cluster.
6. A system according to claim 1 , further comprising:
a cluster presentation module to display each cluster in the set as one of a circular shape and non-circular shape.
7. A system according to claim 1 , further comprising:
a cluster completion module to finalize the clusters in the set prior to generating the cluster spines by performing at least one of merging two or more of the clusters into a single cluster, splitting a single cluster into two or more clusters, and removing minimal clusters.
8. A system according to claim 1 , further comprising:
a calculation module to calculate each of the distances between the cluster and the common origin and between the document and the common origin as a Euclidean distance.
9. A system according to claim 1 , wherein each document in the set appears in at least one cluster.
10. A system according to claim 1 , wherein each cluster in the set comprises a center of mass and defines a convex volume.
11. A computer-implemented method for generating clusters for placement into a display, comprising:
generating a set of clusters from a document set, comprising:
obtaining a single cluster of related documents from the document set; and
adding at least one new cluster, comprising:
comparing one such document in the set to the cluster;
determining a difference in distance between the document and a common origin and the cluster and the common origin; and
designating the document as the new cluster when the difference fails to satisfy a predetermined similarity threshold;
placing into a display one or more cluster spines each comprising two or more of the clusters in the set placed along a vector; and
identifying the clusters along each spine as similar, wherein the clusters in one such spine are also similar to further clusters located along a further spine having a small cosine rotation from that cluster spine.
12. A method according to claim 11 , wherein the display is three-dimensional and each cluster represents multi-dimensional data modeled in the three-dimensional display.
13. A method according to claim 11 , further comprising:
identifying for each cluster in the set a center and radius.
14. A method according to claim 13 , wherein the center of each cluster is located at a fixed distance from the common origin.
15. A method according to claim 13 , wherein the radius of each cluster in the set is independent of the other clusters and reflects a relative number of the related documents in that cluster.
16. A method according to claim 11 , further comprising:
displaying each cluster in the set as one of a circular shape and non-circular shape.
17. A method according to claim 11 , further comprising:
finalizing the clusters in the set prior to generating the cluster spines, comprising at least one of:
merging two or more of the clusters into a single cluster;
splitting a single cluster into two or more clusters; and
removing minimal clusters.
18. A method according to claim 11 , further comprising:
calculating each of the distances as a Euclidean distance.
19. A method according to claim 11 , wherein each document in the set appears in at least one cluster.
20. A method according to claim 11 , wherein each cluster in the set comprises a center of mass and defines a convex volume.