Machine for analyzing entity-resolved data graphs using peer data structures

Abstract

The machine analyzes entity-resolved data graphs using a curated set of peer-reviewed data structures. [Solution] The machine uses independent peer data structures that model similar or related but different universes and entities. These data structures may include other entity-resolved data graphs or file-based data structures. The machine first performs candidate screening to determine which of these data structures meets the requirements for use in the analysis. The machine then selects from the candidate peer data structures by which it will perform the analysis. Finally, the entity-resolved data graph is analyzed using queries against the selected peer data structure to provide an analysis of the quality of that data graph.

Claims

1. A machine for analyzing a subject entity resolution data graph using a peer data structure, wherein the subject entity resolution data graph includes a plurality of asserted relationships, each asserted relationship includes a set of entity data for a particular entity and connections between the entity data, and the machine, At least one processor, The system comprises at least one memory that stores instructions executable by the at least one processor, which facilitates the performance of the operation, Analyzing the relevance of each context of multiple peer entity resolution data structures to generate a set of reliability metrics, wherein the set of reliability metrics is applied to generate multiple candidate data structures, and the multiple candidate data structures form a subset of the multiple peer entity resolution data graphs. The process involves generating combinations of one of the multiple candidate data structures and another of the multiple candidate data structures, analyzing each combination to generate multiple oracles, wherein each of the multiple oracles forms a subset of the multiple candidate data structures, and each of the multiple oracles includes an independent system of peer data for the subject entity resolution data graph. A machine that extracts asserted relationships from each of the plurality of candidate data structures, wherein the asserted relationships consist of sets of data and connections between such data, each of the connected sets of data belongs to a specific entity, and further matches the asserted relationships of these candidate data structures against the subject entity resolution data graph by partitioning the subject entity resolution data graph into disproportionately smaller subsets, thereby generating a set of matching results, each subset being self-contained, the partitioning starting with a set of entities sharing the best address and then expanding to a field that provides a general locality index, and once the partitioning converges, a feedback loop is performed on a set of similarity indexes to define, determining whether a few adjustments to the partition elements are needed, and then performs data-level and entity-level evaluations of the set of matching results, thereby generating a set of quality metrics for the subject entity resolution data graph.

2. The machine according to claim 1, wherein at least one of the plurality of peer entity resolution data structures is a file-based data structure.

3. The machine according to claim 2, wherein at least one of the file-based data structures further comprises a plurality of past versions of at least one of the file-based data structures.

4. The machine according to claim 3, wherein the reliability metrics include source evaluation metrics, match metrics, and evolutionary metrics.

5. The machine according to claim 4, wherein the match matrix is ​​generated by performing a linking process on the subject entity resolution data graph using at least one of the file-based data structures.

6. The machine according to claim 5, wherein the evolutionary metrics are generated by measuring the extent to which the file-based data structure changes over time with respect to at least one of the multiple past versions of the file-based data structure.

7. The machine according to claim 2, wherein the stored executable instructions apply at least one external entity resolution data graph to a subset of internal system data sets to generate a set of match results containing matching data between the at least one external entity resolution data graph and the subset of internal system data sets, and generate a proxy file-based data structure.

8. The machine according to claim 7, wherein at least one of the generated combinations is the proxy file-based data structure combined with at least one of the file-based data structures.

9. The machine according to claim 8, wherein the stored executable instructions further call an external matching service for entity-level evaluation of matching against the proxy file-based data structure.

10. The machine according to claim 2, wherein each of the plurality of peer entity resolution data structures includes specific entity data that is more localized than the subject entity resolution data graph.

11. The machine according to claim 10, wherein each of the plurality of peer entity resolution data structures is independent of each of the other of the plurality of peer entity resolution data structures.

12. The machine according to claim 11, wherein the subject entity resolution data graph includes a plurality of asserted relationships representing the complete universe of entities, and each of the plurality of peer entity resolution data structures includes a subset of the plurality of asserted relationships representing the complete universe of entities.

13. The machine according to claim 2, wherein the independent entity resolution data system is not shareable.

14. The machine according to claim 13, wherein the independent entity resolution data system includes a linking service configured to link data belonging to corresponding entities.

15. A machine for evaluating subject entity resolution data graphs, wherein the machine is A subject entity resolution data graph, wherein the subject entity resolution data graph includes a plurality of asserted relationships, each of which includes at least one touchpoint and at least one identifier, and further includes a linking service configured to receive a touchpoint and return an identifier. Multiple peer data structures, each of which is independent of the other peer data structures, A source evaluation system configured to read a set of asserted relationships from at least one of the plurality of peer data structures, and to use the subject entity resolution data graph to generate a set of source evaluation metrics that demonstrate the consistency of the asserted relationships from the peer data structures, A linking evaluation system configured to read a set of asserted relationships from the peer data structure, compare the set of asserted relationships with the subject entity resolution data graph, and generate a set of match metrics indicating the similarity of the peer data structure to the subject entity resolution data graph, An asserted relationship matching system configured to generate a matching service corresponding to each peer data structure, including a file-based data structure, for each peer data structure, by partitioning the subject entity resolution data graph into disproportionately smaller subsets, wherein each subset is self-contained, the partitioning starts with a set of entities sharing the best address, then extends to a field providing a general locality index, and once the partitioning converges, a feedback loop is performed on the defined set of similarity indexes to determine whether a few adjustments to the partition elements are necessary. A machine comprising: a peer data structure and an evolutionary analysis system configured to read the peer data structure and at least one past version of the peer data structure and generate a set of evolutionary metrics showing how the peer data structure changes over time.

16. The machine according to claim 15, further comprising a quality evaluation system configured to receive data-level evaluations and entity-level evaluations in order to generate a set of quality metrics indicating the degree of similarity between entity data from the peer data structure and entity data from the subject entity resolution data graph.

Images