Pair selection for entity resolution analysis
AssigneeInternational Business Machines Corporation
InventorsAbhishek Seth, Soma Shekar Naganna, Geetha Sravanthi Pulipaty, Rishabh Saraf, Mohit Singh Chouhan
Life of this patent
Estimating the term…
Abstract
Analyzing entity matching systems by determining autolink (AL) and clerical review (CR) threshold values according to record pair compare method scores, defining a score region of interest adjacent to at least one of the AL or CR thresholds, identifying anomalous pairs in at least one region of interest, and providing at least one anomalous record pair to a user for review.
Claims20
A computer implemented method for analyzing entity matching systems, the method comprising:
- determining, by one or more computer processors, an autolink (AL) threshold value according to record pair attribute compare method maximum scores;
- determining, by one or more computer processors, a clerical review (CR) threshold value according to record pair attribute compare method average scores;
- defining, by the one or more computer processors, score regions of interest adjacent to at least one of the AL or CR thresholds;
- identifying, by the one or more computer processors, at least one anomalous record pair in at least one score region of interest;
- and
- providing, by the one or more computer processors, at least one anomalous record pair to a user for review.
A computer program product for analyzing entity matching systems, the computer program product comprising one or more computer readable storage devices and collectively stored program instructions on the one or more computer readable storage devices, the stored program instructions comprising:
- program instructions to determine an autolink (AL) threshold value according to record pair attribute compare method maximum scores;
- program instructions to determine a clerical review (CR) threshold value according to record pair attribute compare method average scores;
- program instructions to define a score regions of interest adjacent to at least one of the AL or CR thresholds;
- program instructions to identify at least one anomalous record pair in at least one score region of interest;
- and
- program instructions to provide at least one anomalous record pair to a user for review.
A computer system for analyzing entity matching systems, the computer system comprising:
- one or more computer processors;
- one or more computer readable storage devices;
- and
- stored program instructions on the one or more computer readable storage devices for execution by the one or more computer processors, the stored program instructions comprising:
- program instructions to determine an autolink (AL) threshold value according to record pair attribute compare method maximum scores;
- program instructions to determine a clerical review (CR) threshold value according to record pair attribute compare method average scores;
- program instructions to define a score regions of interest adjacent to at least one of the AL or CR thresholds;
- program instructions to identify anomalous pairs in at least one score region of interest;
- and
- program instructions to provide at least one anomalous record pair to a user for review.
Dependencies are inferred from the claim text.