System for adaptive multi-cultural searching and matching of personal names

• US 8,041,560 B2
• Filed: 08/22/2008
• Issued: 10/18/2011
• Est. Priority Date: 03/25/1998
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: McCallum-Bayliss, Heather, Shaefer, Leonard Arthur Jr., Lutz, Richard D., Hermansen, John Christian
• Primary Examiner(s): RIDER, JUSTIN W
• Application Number:
  US12/197,119
• Publication Number:
  US 20080312909A1
• Time in Patent Office: 1,152 Days
• Field of Search: 704/10, 704/3, 704/6
• US Class Current: 704/10
• CPC Class Codes: G06F 16/33 Querying
• CPC Class Codes: G06F 16/90344 by using string matching techniques
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99936 Pattern matching access
• CPC Class Codes: Y10S 707/99942 Manipulating data structure, e.g. compression, compaction, compilation
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing

Abstract

An automated name searching system incorporates an automatic name classifier and a multi-path architecture in which different algorithms are applied based on cultural identity of the query name. The name classifier operates with a preemptive list, analysis of morphological elements, length, and linguistic rules. A name regularizer produces a character based computational representation of the name. A pronunciation equivalent representation such as an IPA language representation, and language specific rules to generate name searching keys, are used in a first pass to eliminate database entries which are obviously not matches for the query name. The methods can also be implemented as a callable set of library routines including an intelligent preprocessor and a name evaluator that produces a score comparing a query name and database name, based on a variety of user-adjustable parameters. The user-controlled parameters permit tuning of the search methodologies for specific custom applications.

Claims

1. An apparatus comprising a tangible computer readable storage medium having instructions stored thereon that when executed by a machine result in at least the following:

classifying a text input name as belonging to a particular culture by;

using a high frequency name data store of names that occur frequently in particular cultures, wherein, when there is a match with a name in the high frequency name data store of names, the particular culture associated with retrieved name and a confidence score associated with the retrieved name are recorded,

determining whether morphemes in a morpheme data store are present in the input name by searching for matching substrings of name segments in the input name, and wherein, for each morpheme found in the input name, the particular culture associated with the morpheme and a confidence level associated with the morpheme are recorded, searching the input name for strings of letters that occur with statistical significance in particular cultures, wherein, for each n-gram present in an associated n-gram data store, when a match is found, the culture and score associated with that n-gram are recorded, and

breaking the name into segments and using information in the segments to match at least one of a title, an affix, and a qualifier of the text input name, wherein, for each segment present in the input name that matches a particle in a data store, the culture associated with that particle and a confidence score associated with that particle are recorded;

accessing the text input name entered as an input name by one or more of a user or a system;

determining multiple phonetic representations for a portion of the text input name, each of the multiple phonetic representations being for a different pronunciation of the text input name;

comparing each of the multiple phonetic representations of the portion of the text input name to a phonetic representation of a portion of a text known name stored in a database, wherein comparing each of the multiple phonetic representations of the portion of the text input name to the phonetic representation of the portion of the text known name comprises comparing, for at least one of the multiple phonetic representations of the portion of the text input name, corresponding parts of (i) the at least one phonetic representation of the portion of the text input name and (ii) the phonetic representation of the portion of the text known name, wherein the corresponding parts include parts that correspond at a phonologic level, wherein the parts that correspond at the phonologic level include (i) a first part that relates to a final phoneme of the portion of the text input name and (ii) a second part that relates to a final phoneme of the portion of the text known name; and

providing an indication of whether the text input name matches the text known name based on the comparing.

2. The apparatus of claim 1 wherein the instructions, when executed by a machine, further result in at least the following:

selecting a rule based on the classifying of the text input name; and

applying the rule in determining the multiple phonetic representations for the portion of the text input name.

3. The apparatus of claim 1 wherein the instructions, when executed by a machine, further result in at least the following:

selecting multiple rules based on the classifying of the text input name; and

applying the multiple rules in determining the multiple phonetic representations for the portion of the text input name.

4. The apparatus of claim 1 wherein:

comparing each of the multiple phonetic representations of the portion of the text input name to the phonetic representation of the portion of the text known name comprises determining articulatory similarity between at least one of the multiple phonetic representations of the portion of the text input name and the phonetic representation of the portion of the text known name, and

providing the indication comprises providing an indication of articulatory similarity between the text input name and the text known name, the indication of articulatory similarity being based on the determining of articulatory similarity.

5. The apparatus of claim 4 wherein the instructions, when executed by a machine, further result in at least the following:

identifying an articulatory variation between (i) one or more of the multiple phonetic representations of the portion of the text input name and (ii) the phonetic representation of the portion of the text known name; and

classifying the articulatory variation as likely or unlikely, and

wherein determining articulatory similarity comprises attributing less significance to the articulatory variation, so as to indicate greater articulatory similarity, if the articulatory variation is likely than if the articulatory variation is unlikely.

6. The apparatus of claim 4 wherein determining articulatory similarity comprises determining articulatory similarity based on a culture-specific rule.

7. The apparatus of claim 4 wherein:

determining articulatory similarity between at least one of the multiple phonetic representations of the portion of the text input name and the phonetic representation of the portion of the text known name comprises determining, for the at least one of the multiple phonetic representations of the portion of the text input name, how many phonetic features are in common between corresponding portions of the at least one phonetic representation of the portion of the text input name and the phonetic representation of the portion of the text known name, and

providing the indication of articulatory similarity comprises providing an indication that is based on the determining of how many phonetic features are in common.

8. The apparatus of claim 7 wherein:

the at least one phonetic representation of the portion of the text input name comprises an International Phonetic Alphabet (“

IPA”

) representation of the text input name,

the phonetic representation of the portion of the text known name comprises an IPA representation of the portion of the text known name, and determining how many phonetic features are in common between corresponding portions of the at least one phonetic representation of the portion of the text input name and the phonetic representation of the portion of the text known name comprises determining how many phonetic features are in common between corresponding symbols from the IPA representation of the portion of the text input name and the IPA representation of the portion of the text known name.

9. The apparatus of claim 8 wherein determining how many phonetic features are in common between corresponding symbols from the IPA representation of the portion of the text input name and the IPA representation of the portion of the text known name is based on a culture-specific rule.

10. The apparatus of claim 1 wherein determining multiple phonetic representations comprises determining multiple representations that are each based on an IPA.

11. The apparatus of claim 1 wherein the instructions, when executed by a machine, further result in comparing each of the multiple phonetic representations of the portion of the text input name to a second phonetic representation of the portion of the text known name.

12. The apparatus of claim 1 wherein accessing the text input name comprises accessing a character representation of the text input name.

13. The method of claim 1 wherein the corresponding parts include parts that correspond at a syntactic level.

14. The apparatus of claim 1 wherein the corresponding parts include parts that correspond at a syllabic level.

15. The apparatus of claim 14 wherein the parts that correspond at the syllabic level include (i) a first part that relates to a left-most syllable of the portion of the text input name and (ii) a second part that relates to a left-most syllable of the portion of the text known name.

16. The apparatus of claim 1 wherein comparing each of the multiple phonetic representations of the portion of the text input name to the phonetic representation of the portion of the text known name comprises comparing, for at least one of the multiple phonetic representations of the portion of the text input name, sonority level between at least part of (i) the at least one of the multiple phonetic representations of the portion of the text input name and (ii) the phonetic representation of the portion of the text known name.

17. The apparatus of claim 1 wherein providing the indication of whether the text input name matches the text known name comprises providing a rank-ordered list of names, with rank-order indicating a likelihood of matching the text input name.

18. The apparatus of claim 17 wherein providing the rank-ordered list of names comprises ranking names on the rank-ordered list based on a degree of articulatory similarity between names on the rank-ordered list and the text input name.

19. An apparatus comprising a tangible computer readable storage medium having instructions stored thereon that when executed by a machine result in at least the following:

classifying a text input name as belonging to a particular culture by;

using a high frequency name data store of names that occur frequently in particular cultures, wherein, when there is a match with a name in the high frequency name data store of names, the particular culture associated with retrieved name and a confidence score associated with the retrieved name are recorded,

determining whether morphemes in a morpheme data store are present in the input name by searching for matching substrings of name segments in the input name, and wherein, for each morpheme found in the input name, the particular culture associated with the morpheme and a confidence level associated with the morpheme are recorded,

searching the input name for strings of letters that occur with statistical significance in particular cultures, wherein, for each n-gram present in an associated n-gram data store, when a match is found, the culture and score associated with that n-gram are recorded, and

breaking the name into segments and using information in the segments to match at least one of a title, an affix, and a qualifier of the text input name, wherein, for each segment present in the input name that matches a particle in a data store, the culture associated with that particle and a confidence score associated with that particle are recorded;

accessing the text input name entered as an input name by one or more of a user or a system;

determining multiple phonetic representations for a portion of the text input name, each of the multiple phonetic representations being for a different pronunciation of the text input name;

comparing each of the multiple phonetic representations of the portion of the text input name to a phonetic representation of a portion of a text known name stored in a database, wherein comparing each of the multiple phonetic representations of the portion of the text input name to the phonetic representation of the portion of the text known name comprises comparing, for at least one of the multiple phonetic representations of the portion of the text input name, corresponding parts of (i) the at least one phonetic representation of the portion of the text input name and (ii) the phonetic representation of the portion of the text known name, wherein the corresponding parts include parts that correspond at a syllabic level, wherein the parts that correspond at the syllabic level include (i) a first part that relates to a left-most syllable of the portion of the text input name and (ii) a second part that relates to a left-most syllable of the portion of the text known name, the first part further relates to both an initial phonologic element and a final phonologic element of the left-most syllable of the portion of the text input name, and the second part further relates to an initial phonologic element and a final phonologic element of the left-most syllable of the portion of the text known name; and

providing an indication of whether the text input name matches the text known name based on the comparing.

20. The apparatus of claim 19 wherein the instructions, when executed by a machine, further result in at least the following:

producing a result from the comparing of the first part and the second part; and

determining, based on the result, whether to continue comparing the at least one of the multiple phonetic representations of the portion of the text input name and the phonetic representation of the portion of the text known name.