Efficient data clustering
A data processing technique includes: accessing a matrix (M) representing a graph; wherein: the graph comprises a plurality of nodes to be clustered and a plurality of edges; an edge in the plurality of edges represents an association between two of the plurality of nodes; and an entry of the matrix has a corresponding edge among the plurality of edges. The technique further includes performing an operation on the matrix to generate a result matrix, the operation includes a multiplication function on the matrix; and identifying one or more clusters among the plurality of nodes, based at least in part on the result matrix, including detecting one or more vertices among the plurality of nodes using the result matrix.
1. A method, comprising:
accessing a matrix (M) representing a graph; wherein:
the graph comprises a plurality of nodes and a plurality of edges;
a node in the plurality of nodes represents an object or entity; and
an edge in the plurality of edges represents an association between two of the plurality of nodes;
performing an operation on the matrix to generate a result matrix, the operation includes a multiplication function on the matrix;
identifying one or more clusters among the plurality of nodes, based at least in part on the result matrix, including detecting one or more vertices among the plurality of nodes using the result matrix; and
outputting information about the identified one or more clusters to be displayed, further processed, or both.
2. The method of claim 1 , wherein an entry of the matrix is set according to whether there is an edge between a set of nodes related to the entry.
3. The method of claim 1 , wherein:
the performing of the operation and the identifying of the one or more clusters are a part of processing one or more multi-dimensional tensors on a plurality of processors in parallel.
4. The method of claim 1 , wherein the matrix is an N×N matrix, and wherein N corresponds to number of nodes to be clustered.
5. The method of claim 1 , wherein the multiplication function on the matrix includes M 2 .
6. The method of claim 1 , wherein the multiplication function on the matrix includes M(M−I) j , and wherein I is an identity matrix and j is a positive integer.
7. The method of claim 1 , wherein:
the multiplication function on the matrix includes M(M−I) j ;
I is an identity matrix and j is a positive integer; and
the one or more vertices are one or more j-th order vertices.
8. The method of claim 1 , wherein the one or more vertices among the plurality of nodes are detected by comparing entries of the result matrix with a threshold value or selecting one or more top entries in the result matrix.
9. The method of claim 1 , wherein the plurality of edges are associated with a corresponding plurality of weighted values.
10. The method of claim 1 , wherein at least some of the plurality of edges are associated with negative values.
11. The method of claim 1 , wherein:
at least some of the plurality of edges are associated with negative values;
the one or more vertices among the plurality of nodes are detected by identifying one or more entries of the result matrix that exceed a negative threshold value or selecting one or more bottom entries in the result matrix; and
the one or more clusters are identified as one or more unfriendly clusters.
12. The method of claim 1 , wherein:
at least some of the plurality of edges are associated with negative values;
the one or more vertices among the plurality of nodes are detected by identifying one or more entries of the result matrix that exceed a negative threshold value or selecting one or more bottom entries in the result matrix;
the one or more clusters are identified as one or more unfriendly clusters; and
the method further comprises:
clustering unconnected nodes in an identified unfriendly cluster to be in a friendly cluster.
13. A system, comprising:
one or more processors configured to:
access a matrix (M) representing a graph; wherein:
the graph comprises a plurality of nodes and a plurality of edges;
a node in the plurality of nodes represents an object or entity; and
an edge in the plurality of edges represents an association between two of the plurality of nodes;
perform an operation on the matrix to generate a result matrix, the operation includes a multiplication function on the matrix;
identify one or more clusters among the plurality of nodes, based at least in part on the result matrix, including detecting one or more vertices among the plurality of nodes using the result matrix; and
output information about the identified one or more clusters to be displayed, further processed, or both; and
one or more memories coupled to the one or more processors and configured to provide the one or more processors with instructions.
14. The system of claim 13 , wherein an entry of the matrix is set according to whether there is an edge between a set of nodes related to the entry.
15. The system of claim 13 , wherein:
to perform the operation and to identify the one or more clusters are a part of processing one or more multi-dimensional tensors on a plurality of processors in parallel.
16. The system of claim 13 , wherein the matrix is an N×N matrix, and wherein N corresponds to number of nodes to be clustered.
17. The system of claim 13 , wherein the multiplication function on the matrix includes M 2 .
18. The system of claim 13 , wherein the multiplication function on the matrix includes M(M−I) j , and wherein I is an identity matrix and j is a positive integer.
19. The system of claim 13 , wherein:
the multiplication function on the matrix includes M(M−I) j ;
I is an identity matrix and j is a positive integer; and
the one or more vertices are one or more j-th order vertices.
20. The system of claim 13 , wherein the one or more vertices among the plurality of nodes are detected by comparing entries of the result matrix with a threshold value or selecting one or more top entries in the result matrix.
21. The system of claim 13 , wherein the plurality of edges are associated with a corresponding plurality of weighted values.
22. The system of claim 13 , wherein at least some of the plurality of edges are associated with negative values.
23. The system of claim 13 , wherein:
at least some of the plurality of edges are associated with negative values;
the one or more vertices among the plurality of nodes are detected by identifying one or more entries of the result matrix that exceed a negative threshold value or selecting one or more bottom entries in the result matrix; and
the one or more clusters are identified as one or more unfriendly clusters.
24. The system of claim 13 , wherein:
at least some of the plurality of edges are associated with negative values;
the one or more vertices among the plurality of nodes are detected by identifying one or more entries of the result matrix that exceed a negative threshold value or selecting one or more bottom entries in the result matrix;
the one or more clusters are identified as one or more unfriendly clusters; and
the one or more processors are further configured to:
cluster unconnected nodes in an identified unfriendly cluster to be in a friendly cluster.
25. A computer program product embodied in a tangible non-transitory computer readable storage medium and comprising computer instructions for:
accessing a matrix (M) representing a graph; wherein:
the graph comprises a plurality of nodes and a plurality of edges;
a node in the plurality of nodes represents an object or entity; and
an edge in the plurality of edges represents an association between two of the plurality of nodes;
performing an operation on the matrix to generate a result matrix, the operation includes a multiplication function on the matrix;
identifying one or more clusters among the plurality of nodes, based at least in part on the result matrix, including detecting one or more vertices among the plurality of nodes using the result matrix; and
outputting information about the identified one or more clusters to be displayed, further processed, or both.