Determining an association metric for record attributes associated with cardinalities that are not necessarily the same for training and applying an entity resolution model

1. A method performed by at least one processor, the method comprising:

obtaining pairs of training records for training an entity resolution model, wherein the pairs of training records comprise a first pair of training records, wherein the first pair of training records comprises a first record that indicates a first set of values for a first attribute and a second record that indicates a second set of values for the first attribute;

determining a set of association metrics corresponding to the pairs of training records at least by:

identifying a first value of the first set of values;

determining a first set of individual association metrics corresponding respectively to comparisons between the first value of the first set of values and each value of the second set of values;

executing a first-level reduction operation for the first set of individual association metrics, across the second set of values, to generate a first reduced association metric;

storing the first reduced association metric in a set of reduced association metrics;

identifying a second value of the first set of values;

determining a second set of individual associate metrics corresponding respectively to comparisons between the second value of the first set of values and each value of the second set of values;

executing the first-level reduction operation for the second set of individual association metrics, across the second set of values, to generate a second reduced association metric;

storing the second reduced association metric in the set of reduced association metrics;

excluding the first value and the second value, determining a presence of any more values in the first set of values;

based on determining that no more values are present in the first set of values, executing a second-level reduction operation for the set of reduced association metrics across the first set of values to generate a set of association metrics corresponding to the first pair of training records; and

applying a machine learning algorithm to the set of association metrics corresponding to the pairs of training records to train the entity resolution model.

2. The method of claim 1 , wherein the first attribute comprises a physical location.

3. The method of claim 1 , wherein determining the first set of individual association metrics corresponding respectively to the comparisons between the first value of the first set of values and each of the second set of values comprises:

determining physical distances between the first value and each of the second set of values; and

identifying the physical distances as the first set of individual association metrics.

4. The method of claim 1 , wherein executing the second-level reduction operation for the set of reduced association metrics comprises:

determining a count of any of the set of reduced association metrics that does not satisfy a criteria; and

normalizing the count based on at least one of a first cardinality associated with the first set of values and a second cardinality associated with the second set of values.

5. The method of claim 1 , wherein executing the first-level reduction operation for the first set of individual association metrics comprises at least one of:

determining a minimum value from the first set of individual association metrics;

determining a maximum value from the first set of individual association metrics; and

determining an average value of the first set of individual association metrics.

6. The method of claim 1 , wherein the entity resolution model comprises a random forest model.

7. The method of claim 1 , further comprising:

applying the entity resolution model to target association metrics for a pair of target records to determine a classification of the pair of target records as being associated with a same entity or being associated with different entities.

8. The method of claim 7 , further comprising:

determining that the pair of target records comprises a third record and a fourth record, wherein a first piece of information is indicated by the third record and not indicated by the fourth record; and

responsive to determining that the pair of target records are associated with the same entity, aggregating information indicated by the pair of target records as a set of information associated with the entity, wherein the set of information associated with the entity comprises the first piece of information.

9. The method of claim 8 , further comprising:

receiving a user request for information associated with the entity, wherein the user request does not identify the third record; and

responsive to the user request, presenting, on a user interface, the set of information associated with the entity, the set of information associated with the entity comprising the first piece of information.

10. The method of claim 7 , further comprising:

determining the target association metrics corresponding to the pair of target records at least by:

determining the pair of target records comprises a third record that indicates a third set of values for the first attribute and a fourth record that indicates a fourth set of values for the first attribute;

determining a third set of individual association metrics corresponding respectively to comparisons between a third value of the third set of values and each of the fourth set of values;

determining a fourth set of individual association metrics corresponding respectively to comparisons between a fourth value of the third set of values and each of the second set of values;

executing the first-level reduction operation for the third set of individual association metrics, across the fourth set of values, to generate a third reduced association metric;

executing the first-level reduction operation for the fourth set of individual association metrics, across the fourth set of values, to generate a fourth reduced association metric; and

executing the second-level reduction operation for at least the third reduced association metric and the fourth reduced association metric, across the third set of values, to generate a first target association metric, of the target association metrics, corresponding to the pair of target records.

11. The method of claim 7 , wherein applying the entity resolution model to the target association metrics for the pair of target records to determine the classification of the pair of target records as being associated with a same entity or being associated with different entities comprises:

applying the entity resolution model to the target association metrics for the pair of target records to determine a probability that the pair of target records are associated with the same entity; and

performing one of:

responsive to determining that the probability is above a threshold value, determining the pair of target records as being associated with the same entity; or

responsive to determining that the probability is below the threshold value, determining the pair of target records as being associated with different entities.

12. A computer system comprising:

a hardware processor configured to:

obtain pairs of training records for training an entity resolution model, wherein the pairs of training records comprise a first pair of training records, wherein the first pair of training records comprises a first record that indicates a first set of values for a first attribute and a second record that indicates a second set of values for the first attribute;

determine a set of association metrics corresponding to the pairs of training records at least by being configured to:

identify a first value of the first set of values;

determine a first set of individual association metrics corresponding respectively to comparisons between the first value of the first set of values and each value of the second set of values;

execute a first-level reduction operation for the first set of individual association metrics, across the second set of values, to generate a first reduced association metric;

store the first reduced association metric in a set of reduced association metrics;

identify a second value of the first set of values;

determine a second set of individual associate metrics corresponding respectively to comparisons between the second value of the first set of values and each value of the second set of values;

execute the first-level reduction operation for the second set of individual association metrics, across the second set of values, to generate a second reduced association metric;

store the second reduced association metric in the set of reduced association metrics:

exclude the first value and the second value, determining a presence of any more values in the first set of values;

based on determining that no more values are present in the first set of values, execute a second-level reduction operation for the set of reduced association metrics across the first set of values to generate a set of association metrics corresponding to the first pair of training records; and

apply a machine learning algorithm to the set of association metrics corresponding to the pairs of training records to train the entity resolution model.

13. The computer system of claim 12 , wherein the hardware processor being configured to execute the second-level reduction operation for the set of reduced association metric further comprises the hardware processor being configured to:

determine a count of any of the set of reduced association metrics that does not satisfy a criteria; and

normalize the count based on at least one of a first cardinality associated with the first set of values and a second cardinality associated with the second set of values.

14. The computer system of claim 12 , wherein the hardware processor is further configured to:

apply the entity resolution model to target association metrics for a pair of target records to determine a classification of the pair of target records as being associated with a same entity or being associated with different entities.

15. A method performed by at least one processor, the method comprising:

obtaining pairs of training records for training an entity resolution model, wherein the pairs of training records comprise a first pair of training records, wherein the first pair of training records comprises a first record that indicates a first set of values, including a first value, for a first attribute and a second record that indicates a second set of values, including a second value, for the first attribute;

extracting, from the first set of values, a third set of values for a first portion of the first attribute and a fourth set of values for a second portion of the first attribute, wherein a third value, of the third set of values, and a fourth value, of the fourth set of values, is extracted from the first set of values for the first attribute;

extracting, from the second set of values, a fifth set of values for the first portion of the first attribute and a sixth set of values for the second portion of the first attribute;

determining a set of association metrics corresponding to the pairs of training records at least by:

determining a first set of preliminary association metrics corresponding respectively to comparisons between the third value, of the third set of values, and each value of the fifth set of values;

determining a second set of preliminary association metrics corresponding respectively to comparisons between the fourth value, of the fourth set of values, and each value of the sixth set of values;

based on at least the first set of preliminary association metrics and the second set of preliminary association metrics, determining a first set of individual association metrics corresponding respectively to comparisons between the first value and each value of the second set of values;

executing a first-level reduction operation for the first set of individual association metrics, across the second set of values, to generate a first reduced association metric;

storing the first reduced association metric in a set of association metrics;

executing the first-level reduction operation for a second set of individual association metrics, across the second set of values, to generate a second reduced association metric;

storing the second reduced association metric in the set of association metrics;

determining a presence of any more values in the first set of values;

based on a determination that no more values are present in the first set of values, executing a second-level reduction operation for the set of reduced association metrics to generate a set of association metrics corresponding to the first pair of training records;

applying a machine learning algorithm to the set of association metrics corresponding to the pairs of training records to train the entity resolution model; and

applying the entity resolution model to target association metrics for a pair of target records to determine a classification of the pair of target records as being associated with a same entity or being associated with different entities.

16. The method of claim 15 , wherein:

the first portion of the first attribute comprises a given name attribute; and

the second portion of the first attribute comprises a surname attribute.

17. The method of claim 15 , wherein determining the first set of individual association metrics corresponding respectively to comparisons between the first value and each of the second set of values comprises:

determining weighted averages based on at least the first set of preliminary association metrics and the second set of preliminary association metrics.

18. The method of claim 15 , wherein determining the first set of preliminary association metrics corresponding respectively to comparisons between the third value, of the third set of values, and each of the fifth set of values comprises:

determining Jaro-Winkler distances between the third value and each of the fifth set of values; and

identifying the Jaro-Winkler distances as the first set of preliminary association metrics.

19. The method of claim 15 , wherein executing the second-level reduction operation for the at least the first reduced association metric and the second reduced association metric comprises:

determining a count of any of the at least the first reduced association metric and the second reduced association metric that does not satisfy a criteria; and

normalizing the count based on at least one of a first cardinality associated with the first set of values and a second cardinality associated with the second set of values.

20. The method of claim 15 , wherein executing the first-level reduction operation for the first set of individual association metrics comprises at least one of:

determining a minimum value from the first set of individual association metrics;

determining a maximum value from the first set of individual association metrics; or

determining an average value of the first set of individual association metrics.