Adaptive field-level matching

1. A system for adaptive field-level matching, the system comprising:

one or more processors; and

a non-transitory computer readable medium storing a plurality of instructions, which when executed, cause the one or more processors to:

identify first elements in a field of a prospective record for a database, and second elements in the field of a candidate record, in the database, for matching the prospective record;

identify features corresponding to any of the first elements that are identical to any of the second elements, any of the first elements that are absent from the second elements, and any of the second elements that are absent from the first elements;

determine, by a machine-learning model, a field match score, based on the features, for the second elements in the field of the candidate record;

determine, by another machine-learning model, a record match score, based on weighing the field match score and weighing another field match score for another field of the candidate record, for the candidate record;

determine whether the record match score satisfies a threshold; and

identify the candidate record as matching the prospective record, in response to a determination that the record match score satisfies the threshold.

2. The system of claim 1 , wherein the plurality of instructions further causes the processor to train the machine-learning model to determine field match scores, based on features associated with elements in fields of prospective records and elements in the fields of candidate records, for the elements in the fields of the candidate records.

3. The system of claim 1 , wherein the plurality of instructions further causes the processor to train the other machine-learning model to generate weights for field match scores, and determine record match scores, based on using the weights for weighing the field match scores, for candidate records.

4. The system of claim 1 , wherein the field comprises a postal code field, the first elements and the second elements comprise first characters and second characters, and the features comprise at least one of consecutive characters of the first characters which correspond to and are identical to consecutive characters of the second characters, a suffix comprising at least one of the first characters that lacks correspondence to any of the second characters, a prefix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters, and a suffix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters.

5. The system of claim 1 , wherein the field comprises a phone number field, the first elements and the second elements comprise first characters and second characters, and the features comprise at least one of consecutive characters of the first characters which correspond to and are identical to consecutive characters of the second characters, a prefix comprising at least one of the first characters that lacks correspondence to any of the second characters, a prefix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters, and a suffix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters.

6. The system of claim 1 , wherein the field comprises a personal first name field, the first elements and the second elements comprise first characters and second characters, and the features comprise at least one of consecutive characters of the first characters which correspond to and are identical to consecutive characters of the second characters, a prefix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters, a prefix comprising at least one of the first characters that lacks correspondence to any of the second characters, a suffix comprising at least one of the first characters that corresponds and differs from at least one of the second characters, and a suffix comprising at least one of the first characters that lacks correspondence to any of the second characters.

7. The system of claim 1 , wherein the field comprises an organization name field, the first elements and the second elements comprise first words and second words, and the features comprise at least one of matched words, which comprise each of the first words that are identical to any of the second words, and a probability that each of the matched words is in a list of organization names, and unmatched words, which comprise each of the first words that are absent from the second words and each of the second words that are absent from the first words, and a probability that each of the unmatched words is in the list of organization names.

8. A computer program product comprising computer-readable program code to be executed by one or more processors when retrieved from a non-transitory computer-readable medium, the program code including instructions to:

identify first elements in a field of a prospective record for a database, and second elements in the field of a candidate record, in the database, for matching the prospective record;

identify features corresponding to any of the first elements that are identical to any of the second elements, any of the first elements that are absent from the second elements, and any of the second elements that are absent from the first elements;

determine, by a machine-learning model, a field match score, based on the features, for the second elements in the field of the candidate record;

determine, by another machine-learning model, a record match score, based on weighing the field match score and weighing another field match score for another field of the candidate record, for the candidate record;

determine whether the record match score satisfies a threshold; and

identify the candidate record as matching the prospective record, in response to a determination that the record match score satisfies the threshold.

9. The computer program product of claim 8 , wherein the program code includes further instructions to train the machine-learning model to determine field match scores, based on features associated with elements in fields of prospective records and elements in the fields of candidate records, for the elements in the fields of the candidate records.

10. The computer program product of claim 8 , wherein the program code includes further instructions to train the other machine-learning model to generate weights for field match scores, and determine record match scores, based on using the weights for weighing the field match scores, for candidate records.

11. The computer program product of claim 8 , wherein the field comprises a postal code field, the first elements and the second elements comprise first characters and second characters, and the features comprise at least one of consecutive characters of the first characters which correspond to and are identical to consecutive characters of the second characters, a suffix comprising at least one of the first characters that lacks correspondence to any of the second characters, a prefix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters, and a suffix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters.

12. The computer program product of claim 8 , wherein the field comprises a phone number field, the first elements and the second elements comprise first characters and second characters, and the features comprise at least one of consecutive characters of the first characters which correspond to and are identical to consecutive characters of the second characters, a prefix comprising at least one of the first characters that lacks correspondence to any of the second characters, a prefix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters, and a suffix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters.

13. The computer program product of claim 8 , wherein the field comprises a personal first name field, the first elements and the second elements comprise first characters and second characters, and the features comprise at least one of consecutive characters of the first characters which correspond to and are identical to consecutive characters of the second characters, a prefix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters, a prefix comprising at least one of the first characters that lacks correspondence to any of the second characters, a suffix comprising at least one of the first characters that corresponds and differs from at least one of the second characters, and a suffix comprising at least one of the first characters that lacks correspondence to any of the second characters.

14. The computer program product of claim 8 , wherein the field comprises an organization name field, the first elements and the second elements comprise first words and second words, and the features comprise at least one of matched words, which comprise each of the first words that are identical to any of the second words, and a probability that each of the matched words is in a list of organization names, and unmatched words, which comprise each of the first words that are absent from the second words and each of the second words that are absent from the first words, and a probability that each of the unmatched words is in the list of organization names.

15. A computer-implemented method for adaptive field-level matching, the computer-implemented method comprising:

identifying first elements in a field of a prospective record for a database, and second elements in the field of a candidate record, in the database, for matching the prospective record;

identifying features corresponding to any of the first elements that are identical to any of the second elements, any of the first elements that are absent from the second elements, and any of the second elements that are absent from the first elements;

determining, by a machine-learning model, a field match score, based on the features, for the second elements in the field of the candidate record;

determining, by another machine-learning model, a record match score, based on weighing the field match score and weighing another field match score for another field of the candidate record, for the candidate record;

determining whether the record match score satisfies a threshold; and

identifying the candidate record as matching the prospective record, in response to a determination that the record match score satisfies the threshold.

16. The computer-implemented method of claim 15 , wherein the computer-implemented method further comprises:

training the machine-learning model to determine field match scores, based on features associated with elements in fields of prospective records and elements in the fields of candidate records, for the elements in the fields of the candidate records; and

training the other machine-learning model to generate weights for the field match scores, and determine record match scores, based on using the weights for weighing the field match scores, for candidate records.

17. The computer-implemented method of claim 15 , wherein the field comprises a postal code field, the first elements and the second elements comprise first characters and second characters, and the features comprise at least one of consecutive characters of the first characters which correspond to and are identical to consecutive characters of the second characters, a suffix comprising at least one of the first characters that lacks correspondence to any of the second characters, a prefix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters, and a suffix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters.

18. The computer-implemented method of claim 15 , wherein the field comprises a phone number field, the first elements and the second elements comprise first characters and second characters, and the features comprise at least one of consecutive characters of the first characters which correspond to and are identical to consecutive characters of the second characters, a prefix comprising at least one of the first characters that lacks correspondence to any of the second characters, a prefix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters, and a suffix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters.

19. The computer-implemented method of claim 15 , wherein the field comprises a personal first name field, the first elements and the second elements comprise first characters and second characters, and the features comprise at least one of consecutive characters of the first characters which correspond to and are identical to consecutive characters of the second characters, a prefix comprising at least one of the first characters that corresponds to and differs from at least one of the second characters, a prefix comprising at least one of the first characters that lacks correspondence to any of the second characters, a suffix comprising at least one of the first characters that corresponds and differs from at least one of the second characters, and a suffix comprising at least one of the first characters that lacks correspondence to any of the second characters.

20. The computer-implemented method of claim 15 , wherein the field comprises an organization name field, the first elements and the second elements comprise first words and second words, and the features comprise at least one of matched words, which comprise each of the first words that are identical to any of the second words, and a probability that each of the matched words is in a list of organization names, and unmatched words, which comprise each of the first words that are absent from the second words and each of the second words that are absent from the first words, and a probability that each of the unmatched words is in the list of organization names.