System for identifying duplicate parties using entity resolution

An entity resolution system comprising a processing device and a memory comprising instructions which are executed by the processing device for performing entity resolution on a data set received by the entity resolution, the data set comprising a plurality of candidate entities and associated entity features, the entity resolution system further comprising:

• a machine learning module configured to generate a synthesized data set, the synthesized data set comprising similarity ratings for each entity feature in the data set, the similarity ratings between two entities of the plurality of candidate entities collectively being a relationship marker associated with those two entities;
• and
• a narrative module configured to:
• determine one or more distances by performing a clustering analysis on the similarity ratings, the one or more distances being measured from each relationship marker to one or more clusters associated with known relationships between entities;
• generate a narrative output based on the one or more distances, the narrative output comprising:
• at least one identified relationship between the first entity and the second entity, and
• a confidence score;
• and
• provide the narrative output to a user interface.

A tangible, non-transitory computer program product configured to perform an entity resolution process, comprising:

• receiving a data set comprising first entity features describing a first entity and second entity features describing a second entity;
• generating a synthesized data set from the first entity features and the second entity features, the synthesized data set comprising similarity ratings for each entity feature of the first and second entity features, the similarity ratings between the first entity and the second entity collectively being a relationship marker;
• determine one or more distances by performing a clustering analysis on the similarity ratings, the one or more distances being measured from the relationship marker to one or more clusters associated with known relationships between entities;
• generating a narrative output based on the one or more distances, the narrative output comprising:
• at least one identified relationship between the first entity and the second entity, and
• a confidence score;
• and
• providing the narrative output to a user interface.