Bipartite graph construction

1. A non-transitory computer readable storage medium having an optimized search engine executed by a hardware processor reading a series of instructions from a hardware memory, comprising:

a database interface configured to access an incoming data item set and a target data item set;

a streamed clustering module configured to group the incoming data item set into incoming clusters and to group the target data item set into target clusters;

a qualification level estimate module configured to generate a cluster qualification level estimate between an incoming cluster of the incoming clusters and a target cluster of the target clusters;

a graph generator configured to build a bipartite graph between the incoming clusters and the target clusters based on the cluster qualification level estimate;

a client input configured to receive an incoming data item from a client device to be assigned to the incoming cluster;

an edge reader configured to consult the bipartite graph to calculate at least one individual qualification level estimate between the incoming data item and at least one target data item;

a report generator configured to generate a report ranking target data item options based on the at least one individual qualification level estimate; and

a client output configured to send the report to the client device.

2. The non-transitory computer readable storage medium having the optimized search engine of claim 1 , wherein the streamed clustering module is configured to identify an incoming cluster center for the incoming cluster.

3. The non-transitory computer readable storage medium having the optimized search engine of claim 2 , wherein the streamed clustering module is configured to compare the incoming data item to the incoming cluster center.

4. The non-transitory computer readable storage medium having the optimized search engine of claim 3 , wherein the streamed clustering module is configured to assign the incoming data item to the incoming cluster based upon an incoming center distance between the incoming data item and the incoming cluster center.

5. The non-transitory computer readable storage medium having the optimized search engine of claim 2 , wherein the qualification level estimate module is configured to generate the cluster qualification level estimate between the incoming cluster center and a target cluster center of the target cluster.

6. The non-transitory computer readable storage medium having the optimized search engine of claim 5 , wherein the qualification level estimate module is configured to generate the at least one individual qualification level estimate by multiplying a cluster qualification level estimate by a target center distance between the target data item and the target cluster center.

7. The non-transitory computer readable storage medium having the optimized search engine of claim 1 , wherein the report generator is configured to generate a follow up report ranking target data item options triggered by an addition of a new target data item.

8. A computing device, having a memory to store a series of instructions that are executed by at least one processor to implement an optimized search engine, the computing device configured to

receive an incoming data item from a client device describing aspects of a client user for matching purposes;

generate an incoming data object to represent the incoming data item;

identify a target cluster of a target object data set representing a target data item set;

read a cluster qualification level estimate for the target cluster in relation to the incoming data object;

calculate an individual qualification level estimate between the incoming data object and a target data object in the target cluster based on the cluster qualification level estimate;

generate a report ranking target data item options based on the individual qualification estimate; and

send the report to the client device.

9. The computing device of claim 8 , wherein the computing device is further configured to

generate a new incoming cluster for the incoming data object when unassigned to an existing incoming cluster.

10. The computing device of claim 8 , wherein the computing device is further configured to

compare the incoming data object to an incoming cluster center for an incoming cluster of an incoming data object set.

11. The computing device of claim 10 , wherein the computing device is further configured to

assign the incoming data object to the incoming cluster based upon an incoming center distance between the incoming data object and the incoming cluster center.

12. The computing device of claim 11 , wherein the computing device is further configured to

determine whether the incoming data object belongs to a new incoming cluster generated from the incoming cluster.

13. The computing device of claim 12 , wherein the computing device is further configured to

set a tuning variable to calculate whether to create a new incoming cluster generated from the incoming cluster.

14. The computing device of claim 11 , wherein the computing device is further configured to

merge the incoming cluster with a nearby incoming cluster.

15. The computing device of claim 14 , wherein the computing device is further configured to

set a merging variable to calculate whether to merge the incoming cluster with the nearby incoming cluster.

16. The computing device of claim 14 , wherein the computing device is further configured to

set a merger probability to equal a ratio of contentious data objects between two incoming clusters to total incoming data objects of the two incoming clusters.

17. A machine-implemented method, comprising:

receiving an incoming data item from a client device describing aspects of a client user for matching purposes;

generating an incoming data object to represent the incoming data item as a member of an incoming data object set representing an incoming data item set;

comparing the incoming data object to an incoming cluster center for an incoming cluster representing of the incoming data object set;

assigning the incoming data object to the incoming cluster based upon an incoming center distance between the incoming data object and the incoming cluster center;

identifying a target cluster of a target object data set representing a target data item set;

reading a cluster qualification level estimate between an incoming cluster center of the incoming cluster and a target cluster center of the target cluster;

calculating an individual qualification level estimate between the incoming data object and a target data object in the target cluster by multiplying the cluster qualification level estimate by a target center distance between the target data object and the target cluster center;

generating a report ranking target data item options based on the individual qualification estimate; and

sending the report to the client device.

18. The method of claim 17 , further comprising:

calculating each incoming vector of the incoming data object in separate category calculations.

19. The method of claim 17 , further comprising:

assigning a standardized score as a category score for an incoming vector.

20. The method of claim 17 , further comprising:

appending a related category score to a category score for an incoming vector.