Candidate data record prioritization for match processing

A computer system comprising:

• a memory;
• and
• a processor in communication with the memory, wherein the computer system is configured to perform a method comprising:
• identifying candidate data records to be used in match processing, the identified candidate data records being from a dataset of data records representing different entities, and the match processing being to determine any of the identified candidate data records, that represent an entity that is a same entity represented by an incoming data record, and therefore that match to the incoming data record, for consolidation of data as between the incoming data record and matching data records of the identified candidate data records to form master data that represents that entity, the identified candidate data records being grouped into a plurality of buckets according to bucket roles, wherein each bucket correlates to a bucket role of the bucket roles, and wherein each bucket role of the bucket roles is defined by a unique record attribute set;
• obtaining effectiveness scores for the bucket roles, the effectiveness scores being measures of effectiveness of the bucket roles in identifying data records that match to incoming data records;
• using a machine learning model to determine data record attributes contributing most to the effectiveness of the bucket roles in identifying data records that match to incoming data records;
• establishing an order of priority in which to process the identified candidate data records in the match processing, the establishing the order of priority comprising prioritizing the plurality of buckets into an order based on the effectiveness scores of the bucket roles, wherein the plurality of buckets are prioritized in an order of descending effectiveness of their correlated bucket roles;
• commencing the match processing to process the identified candidate data records in the established order of priority for processing the identified candidate data records, in which the match processing processes candidate data records, of the identified candidate data records, of a higher priority bucket before processing candidate data records, of the identified candidate data records, of a lower priority bucket;
• based on determining one or more data records of the identified candidate data records that match to the incoming data record, consolidating data of the incoming data record and data of the one or more data records matching to the incoming data record to produce and store a master data record representing that entity;
• and
• managing the bucket roles based on feedback as to which bucket roles provide data records that match to the incoming data records, the managing comprising further training the machine learning model, using the feedback to determine data record attributes contributing most to the effectiveness of the bucket roles, and updating the bucket roles based on the further training and the determined data record attributes contributing most to the effectiveness of the bucket roles.

A computer program product comprising:

• a non-transitory computer readable storage medium readable by a processing circuit and storing instructions for execution by the processing circuit for performing a method comprising:
• identifying candidate data records to be used in match processing, the identified candidate data records being from a dataset of data records representing different entities, and the match processing being to determine any of the identified candidate data records that represent an entity that is a same entity represented by an incoming data record, and therefore that match to the incoming data record, for consolidation of data as between the incoming data record and matching data records of the identified candidate data records to form master data that represents that entity, the identified candidate data records being grouped into a plurality of buckets according to bucket roles, wherein each bucket correlates to a bucket role of the bucket roles, and wherein each bucket role of the bucket roles is defined by a unique record attribute set;
• obtaining effectiveness scores for the bucket roles, the effectiveness scores being measures of effectiveness of the bucket roles in identifying data records that match to incoming data records;
• using a machine learning model to determine data record attributes contributing most to the effectiveness of the bucket roles in identifying data records that match to incoming data records;
• establishing an order of priority in which to process the identified candidate data records in the match processing, the establishing the order of priority comprising prioritizing the plurality of buckets into an order based on the effectiveness scores of the bucket roles, wherein the plurality of buckets are prioritized in an order of descending effectiveness of their correlated bucket roles;
• commencing the match processing to process the identified candidate data records in the established order of priority for processing the identified candidate data records, in which the match processing processes candidate data records, of the identified candidate data records, of a higher priority bucket before processing candidate data records, of the identified candidate data records, of a lower priority bucket;
• based on determining one or more data records of the identified candidate data records that match to the incoming data record, consolidating data of the incoming data record and data of the one or more data records matching to the incoming data record to produce and store a master data record representing that entity;
• and
• managing the bucket roles based on feedback as to which bucket roles provide data records that match to the incoming data records, the managing comprising further training the machine learning model, using the feedback to determine data record attributes contributing most to the effectiveness of the bucket roles, and updating the bucket roles based on the further training and the determined data record attributes contributing most to the effectiveness of the bucket roles.