Systems and methods for interactive clustering
View Patent ↗Systems and associated methods provide a cluster-level semi-supervision model for inter-active clustering. Embodiments accept user provided semi-supervision for updating cluster descriptions and assignment of data items to clusters. Assignment feedback re-assigns data items among existing clusters, while cluster description feedback helps to position existing cluster centers more meaningfully. The feedback can continue until the user is satisfied with the clustering achieved or one or more predetermined stopping criteria have been reached.
1. A computer program product for interactive clustering comprising:
a computer readable storage medium having computer readable program code embodied therewith, the computer readable program code comprising:
computer readable program code configured to access a collection of items to be clustered;
computer readable program code configured to cluster the collection of items to be clustered into an initial plurality of clusters;
computer readable program code configured to access user input feedback modifying one or more of cluster assignment of the initial plurality of clusters and cluster description of the initial plurality of clusters; and
computer readable program code configured to re-cluster the collection of items based on the user input feedback; and
wherein the computer readable program code configured to re-cluster the collection of items based on the user input feedback is further configured to iteratively re-cluster responsive to user feedback.
2. The computer program product according to claim 1 , wherein the user input feedback includes both cluster assignment feedback and cluster description feedback.
3. The computer program product according to claim 2 , wherein the cluster assignment feedback comprises feedback re-assigning one or more clustered objects to a new cluster of the initial plurality of clusters.
4. The computer program product according to claim 2 , wherein the cluster description feedback comprises feedback changing a description of one or more clusters of the initial plurality of clusters.
5. The computer program product according to claim 2 , wherein the collection of items to be clustered comprises a plurality of documents.
6. The computer program product according to claim 5 , wherein the cluster assignment feedback includes one or more indications for re-assigning one or more documents to one or more clusters of the initial plurality of clusters.
7. The computer program product according to claim 5 , wherein the cluster description feedback includes one or more changes to one or more descriptive words used to describe one or more clusters of the initial plurality of clusters.
8. The computer program product according to claim 1 , wherein to cluster the collection of items to be clustered into an initial plurality of clusters further comprises utilizing k-means clustering.
9. A computer-implemented method for interactive clustering comprising:
accessing a collection of items to be clustered;
clustering the collection of items to be clustered into an initial plurality of clusters;
accessing user input feedback modifying one or more of cluster assignment of the initial plurality of clusters and cluster description of the initial plurality of clusters; and
re-clustering the collection of items based on the user input feedback; and
wherein re-clustering the collection of items based on the user input feedback further comprises iteratively re-clustering responsive to user feedback.
10. The method according to claim 9 , wherein the user input feedback includes both cluster assignment feedback and cluster description feedback.
11. The method according to claim 10 , wherein the cluster assignment feedback comprises feedback re-assigning one or more clustered objects to a new cluster of the initial plurality of clusters.
12. The method according to claim 10 , wherein the cluster description feedback comprises feedback changing a description of one or more clusters of the initial plurality of clusters.
13. The method according to claim 10 , wherein the collection of items to be clustered comprises a plurality of documents.
14. The method according to claim 13 , wherein the cluster assignment feedback includes one or more indications for re-assigning one or more documents to one or more clusters of the initial plurality of clusters.
15. The method according to claim 13 , wherein the cluster description feedback includes one or more changes to one or more descriptive words used to describe one or more clusters of the initial plurality of clusters.
16. The method according to claim 9 , wherein clustering the collection of items to be clustered into an initial plurality of clusters further comprises utilizing k-means clustering.
17. A system for interactive clustering comprising:
one or more processors; and
a memory operatively connected to the one or more processors;
wherein, responsive to execution of computer readable program code accessible to the one or more processors, the one or more processors are configured to:
access a collection of items to be clustered;
cluster the collection of items to be clustered into an initial plurality of clusters;
access user input feedback modifying one or more of cluster assignment of the initial plurality of clusters and cluster description of the initial plurality of clusters; and
re-cluster the collection of items based on the user input feedback; and
wherein re-clustering the collection of items based on the user input feedback further comprises iteratively re-clustering responsive to user feedback.
18. The system according to claim 17 , wherein:
feedback modifying cluster assignment comprises feedback re-assigning one or more clustered objects to a new cluster of the initial plurality of clusters; and
feedback modifying cluster description comprises feedback changing a description of one or more clusters of the initial plurality of clusters.