Back to results
US 2008/0189279 A1Published application
Fast accurate fuzzy matching
AssigneeFair Issac Corporation
InventorsUwe MAYER, Vijay Narayanan, Matthias Blume
Life of this patent
Estimating the term…
Abstract
A computer-implemented technique for fuzzy matching. This works quickly yet accurately to determine if a given computer-readable record is represented, by exact match or pretty close match, in a large collection of computer-readable records. Further tools may be provided to assess the character of the match.
Claims20
A method of processing computer-readable data, comprising computer-implemented operations of:
- identifying a candidate pool of records of a multi-record computer-readable reference database by performing fuzzy matching between a first set of fields of an input record and corresponding fields of records from the reference database;
- filtering the candidate pool by removing records from the candidate pool demonstrating prescribed weaknesses of said fuzzy match with the first set of fields of the input record;
- analyzing the filtered candidate pool by performing fuzzy matching between a second set of fields of the input record and corresponding fields of records of the filtered candidate pool;
- performing at least one of the following operations:
- pruning the candidate pool according to results of the analyzing operation and providing a computer-readable output including identification of records of the pruned candidate pool;
- providing a computer-readable output including identification of records of the candidate pool and results of the analyzing operation.
3 more elements
A method of processing computer-readable data, comprising computer-implemented steps of:
- a step for identifying a candidate pool of records of a multi-record computer-readable reference database by performing fuzzy matching between a first set of fields of an input record and corresponding fields of records from the reference database;
- a step for filtering the candidate pool by removing records from the candidate pool demonstrating prescribed weaknesses of said fuzzy match with the first set of fields of the input record;
- a step for analyzing the filtered candidate pool by performing fuzzy matching between a second set of fields of the input record and corresponding fields of records of the filtered candidate pool;
- performing at least one of the following operations:
- a step for pruning the candidate pool according to results of the analyzing operation and providing a computer-readable output including identification of records of the pruned candidate pool;
- a step for providing a computer-readable output including identification of records of the candidate pool and results of the analyzing operation.
3 more elements
A computer readable medium storing a program for perform operations comprising:
- identifying a candidate pool of records of a multi-record computer-readable reference database by performing fuzzy matching between a first set of fields of an input record and corresponding fields of records from the reference database;
- filtering the candidate pool by removing records from the candidate pool demonstrating prescribed weaknesses of said fuzzy match with the first set of fields of the input record;
- analyzing the filtered candidate pool by performing fuzzy matching between a second set of fields of the input record and corresponding fields of records of the filtered candidate pool;
- performing at least one of the following operations:
- pruning the candidate pool according to results of the analyzing operation and providing a computer-readable output including identification of records of the pruned candidate pool;
- providing a computer-readable output including identification of records of the candidate pool and results of the analyzing operation.
3 more elements
A digital data processing apparatus programmed to perform operations comprising:
- identifying a candidate pool of records of a multi-record computer-readable reference database by performing fuzzy matching between a first set of fields of an input record and corresponding fields of records from the reference database;
- filtering the candidate pool by removing records from the candidate pool demonstrating prescribed weaknesses of said fuzzy match with the first set of fields of the input record;
- analyzing the filtered candidate pool by performing fuzzy matching between a second set of fields of the input record and corresponding fields of records of the filtered candidate pool;
- performing at least one of the following operations:
- pruning the candidate pool according to results of the analyzing operation and providing a computer-readable output including identification of records of the pruned candidate pool;
- providing a computer-readable output including identification of records of the candidate pool and results of the analyzing operation.
3 more elements
A computer-implemented method comprising operations of:
- receiving a computer-readable multi-field input record for analysis;
- performing first stage fuzzy matching on limited fields between the computer-readable input record and computer-readable multi-field records in a reference database, the fuzzy matching employing computer-readable keys pre-computed for fields or collections of fields of the reference database;
- filtering matching records according to predetermined statistical criteria;
- performing second stage fuzzy matching as between the input record and the filtered matching records using one of the following:
- data from records of the reference database, the pre-computed keys;
- where the second stage fuzzy matching differs from the first stage fuzzy matching in at least one of the following respects:
- the second stage fuzzy matching is applied to a substantially greater number of fields than the limited fields;
- the second stage fuzzy matching uses a different formula than used in the first stage fuzzy matching;
- performing at least one of the following operations:
- providing a computer-readable output including identification of records resulting from the second stage fuzzy matching;
- providing a computer-readable output including identification of the filtered matching records and an associated rating according to results of the second stage fuzzy matching.
8 more elements
A method of processing computer-readable data, comprising computer-implemented steps of:
- a step for receiving a computer-readable multi-field input record for analysis;
- a step for performing first stage fuzzy matching on limited fields between the computer-readable input record and computer-readable multi-field records in a reference database, the fuzzy matching employing computer-readable keys pre-computed for fields or collections of fields of the reference database;
- a step for filtering matching records according to predetermined statistical criteria;
- a step for performing second stage fuzzy matching as between the input record and the filtered matching records using one of the following:
- data from records of the reference database, the pre-computed keys;
- where the second stage fuzzy matching differs from the first stage fuzzy matching in at least one of the following respects:
- the second stage fuzzy matching is applied to a substantially greater number of fields than the limited fields;
- the second stage fuzzy matching uses a different formula than used in the first stage fuzzy matching;
- performing at least one of the following steps:
- a step for providing a computer-readable output including identification of records resulting from the second stage fuzzy matching;
- a step for providing a computer-readable output including identification of the filtered matching records and an associated rating according to results of the second stage fuzzy matching.
8 more elements
A computer readable medium storing a program for perform operations comprising:
- receiving a computer-readable multi-field input record for analysis;
- performing first stage fuzzy matching on limited fields between the computer-readable input record and computer-readable multi-field records in a reference database, the fuzzy matching employing computer-readable keys pre-computed for fields or collections of fields of the reference database;
- filtering matching records according to predetermined statistical criteria;
- performing second stage fuzzy matching as between the input record and the filtered matching records using one of the following:
- data from records of the reference database, the pre-computed keys;
- where the second stage fuzzy matching differs from the first stage fuzzy matching in at least one of the following respects:
- the second stage fuzzy matching is applied to a substantially greater number of fields than the limited fields;
- the second stage fuzzy matching uses a different formula than used in the first stage fuzzy matching;
- performing at least one of the following operations:
- providing a computer-readable output including identification of records resulting from the second stage fuzzy matching;
- providing a computer-readable output including identification of the filtered matching records and an associated rating according to results of the second stage fuzzy matching.
8 more elements
A digital data processing apparatus programmed to perform operations comprising:
- receiving a computer-readable multi-field input record for analysis;
- performing first stage fuzzy matching on limited fields between the computer-readable input record and computer-readable multi-field records in a reference database, the fuzzy matching employing computer-readable keys pre-computed for fields or collections of fields of the reference database;
- filtering matching records according to predetermined statistical criteria;
- performing second stage fuzzy matching as between the input record and the filtered matching records using one of the following:
- data from records of the reference database, the pre-computed keys;
- where the second stage fuzzy matching differs from the first stage fuzzy matching in at least one of the following respects:
- the second stage fuzzy matching is applied to a substantially greater number of fields than the limited fields;
- the second stage fuzzy matching uses a different formula than used in the first stage fuzzy matching;
- performing at least one of the following operations:
- providing a computer-readable output including identification of records resulting from the second stage fuzzy matching;
- providing a computer-readable output including identification of the filtered matching records and an associated rating according to results of the second stage fuzzy matching.
8 more elements
Dependencies are inferred from the claim text.