Scoring unfielded personal names without prior parsing

• US 10,229,112 B2
• Filed: 03/07/2016
• Issued: 03/12/2019
• Est. Priority Date: 04/13/2015
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Patman Maguire, Frankie E.
• Primary Examiner(s): Kuddus, Daniel A
• Application Number:
  US15/062,281
• Publication Number:
  US 20160299895A1
• Time in Patent Office: 1,100 Days
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 40/295 Named entity recognition
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06F 40/40 Processing or translation of natural language natural language analysis G06F40/20; semantic analysis G06F40/30

Abstract

A system for determining a similarity between a name phrase and a comparison name phrase, for each name in the name phrase, scores the name. The scoring is based on the field frequency of the name in a name database, where the field frequency indicates a given name frequency and/or a surname frequency in the database. The system uses the scoring to determine a transition from a given name to a surname in the name phrase. The system determines a primary given name and a primary surname in the name phrase based on the scoring and the transition. The system uses the primary given name and primary surname to determine a similarity between the name phrase and a comparison name phrase, where the comparison name phrase comprises a comparison given name and a comparison surname.

Claims

1. A method of determining a similarity between a name phrase and a comparison name phrase, the method comprising:

for each name in the name phrase, scoring the name by a computing processor, wherein the scoring is based on field frequency of the name in a name database, wherein the field frequency indicates at least one of a given name frequency and a surname frequency in the database;

using the scoring, by the computing processor, to determine a transition from a given name to a surname in the name phrase, comprising;

calculating, by the computing processor, a transition score for each name in the name phrase for which a preceding name exists, wherein the transition score is calculated using the name and the preceding name, wherein the preceding name precedes the name in an order of names in the name phrase; and

using the transition score, by the computing processor, to determine the primary surname in the name phrase, wherein the transition score indicates a transition from given names in the name phrase to the primary surname in the name phrase;

determining, by the computing processor, a primary given name in the name phrase based on the scoring and the transition;

determining, by the computing processor, a primary surname in the name phrase based on the scoring and the transition; and

using the primary given name and primary surname, by the computing processor, to determine a similarity between the name phrase and a comparison name phrase, wherein the comparison name phrase comprises a comparison given name and a comparison surname.

2. The method of claim 1 wherein for each name in the name phrase, scoring the name, wherein the scoring is based on the field frequency of the name in the name database comprises:

performing the scoring without parsing the name phrase.

3. The method of claim 1 wherein for each name in the name phrase, scoring the name, wherein the scoring is based on the field frequency of the name in the name database comprises:

determining the field frequency according to a country associated with the names in the name database.

4. The method of claim 1 wherein for each name in the name phrase, scoring the name, wherein the scoring is based on the field frequency of the name in the name database comprises:

averaging the field frequency of each name in the name phrase to determine whether an order of names in the name phrase is correct.

5. The method of claim 1, wherein calculating the transition score for each name in the name phrase for which the preceding name exists comprises:

averaging the field frequency of the name and the field frequency of the preceding name to determine the transition score.

6. The method of claim 1 wherein using the primary given name and primary surname to determine the similarity between the name phrase and the comparison name phrase comprises:

using the field frequency to adjust an assessment of similar names, wherein the field frequency indicates whether similar names are likely given names or surnames.

7. The method of claim 1 wherein the scoring of the name, the using of the scoring to determine the transition, the determining of the primary given name, and the determining of the primary surname occur without parsing of the name phrase and without requiring memory to store parsed data.