Elamur Research⌕
☰
home / papers / 2013 / Cross-Language Personal Name Mapping

Cross-Language Personal Name Mapping

Paper · 2013 · Ahmed H. Yousef

Summary

The paper addresses linking person records across databases when the same person's name is stored in Arabic in one source and in English transliteration in another. Arabic names are hard to match. Spellings vary, dialects and pronunciation differ, compound names such as Abdel-Hamid get split or shortened, and name order carries meaning. Yousef builds a dictionary-driven mapping framework with three parts. The first is a revised Arabic Soundex ("combined" Soundex) that recognises Arabic name features and is used to build a bilingual names dictionary automatically. The second is a proximity matcher called Weighted Atomic Token, which respects token order. The third is a set of new quality metrics that avoid being swamped by true negatives. The framework is tested on real Egyptian university data from the ICTP and EUL projects, including Scopus/ISI scholar records matched against the Arabic UMIS system. Dictionaries built with combined Soundex roughly double true-positive rates over dictionaries taken directly from entered data. The framework also handles initials, where edit distance breaks down. This matters for cross-language entity linking and for data integration involving non-Latin scripts.

Key points

  • Record pairs fall into three groups: matched (M), unmatched (U) and possible (P). Pairs in P go to a domain expert for clerical review.
  • Dictionary-building comparison, measured as one-to-one true positives / "Overall True Positive Accuracy Percentage":
    • Raw source dataset: 40.55% / 48.26%
    • Soundex-filtered: 44.62% / 49.60%
    • Combined Arabic Soundex: 69.86% / 77.81%
    • Combined Soundex plus expert verification: 72.01% / 80.37%
  • In the scholar experiment, 1,000 transliterated names, often written with initials like "A. M. Aly", were matched against 80,000 Arabic UMIS records. Results were about 62% true positives and 18% verified true negatives, with original precision of 96% and recall of 93%. Overall, roughly 90% of the names were mapped correctly.
  • Levenshtein distance is ruled out for initials, since "Mohamed" vs "M" already gives a distance of 6.
  • Accuracy is criticised as a metric because the huge number of true negatives dominates it. New percentage-based measures, including an extended true-positive measure, replace it.

Interesting ideas

  • I1. A three-way decision of match, non-match or "possible" sends only the uncertain middle band to a human, so expert effort stays targeted and auditable.
  • I2. A phonetic code (Soundex) serves as a cheap consistency check on dictionary entries, filtering out data-entry errors before the dictionary is trusted.
  • I3. Names are compared token by token, in order, and a token counts as matching if it equals the other token or is its prefix. This lets initials line up with full names where character edit distance fails.
  • I4. Quality metrics are designed so the large count of true negatives cannot inflate the score, focusing on how many real matches are recovered.
  • I5. The authors suggest a cascade of strict matching, then relaxed conditions, then edit distance as a final tie-breaker, so the cheapest reliable rule runs first.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

  • I1 (learn): When the witness parsers disagree, a "possible" bucket could route just those sentences to the LLM teacher tier instead of forcing a silver label.
  • I2 (world): When compiling entity aliases from Wikidata, a phonetic agreement check could flag suspicious alias/transliteration pairs before they enter the knowledge graph.
  • I3 (frames): Coreference could use ordered prefix-token matching to link "J. Smith" or "Mr. Smith" to an earlier "John Smith" with an explicit, explainable reason.
  • I4 (learn): Evaluations with negative controls should report rates that are not dominated by trivially correct negatives, for example in coreference or entity-linking checks.
  • I5 (world): Entity resolution in the world of a text could try exact, then ordered-token, then fuzzy matching, logging which stage made each link.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

📚 Shelves 7The field of science this work belongs to (OpenAlex fields).Computer Science · 9,085 worksArtificial Intelligence · 8,888 worksArts and Humanities · 9,296 worksLanguage and Linguistics · 9,206 worksComputer Vision and Pattern Recognition · 1,282 worksDecision Sciences · 85 worksManagement Science and Operations Research · 36 works

🧭 Topics 2Research subjects and languages the work is about.Multilingual & cross-lingual NLP · 465 worksSemitic languages · 1,485 works

Cite

Cite as Ahmed H. Yousef. 2013. Cross-Language Personal Name Mapping. arXiv:1405.6293 · International Journal of Computational Linguistics Research, vol 4, issue 4, December 2013.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.1405.6293 · arXiv 1405.6293

@misc{yousef2013cross,
  title = {{Cross-Language Personal Name Mapping}},
  author = {Ahmed H. Yousef},
  year = {2013},
  doi = {10.48550/arxiv.1405.6293},
  eprint = {1405.6293},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.1405.6293}
}

References

43 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Paul McNamee, J.M., Dawn Lawrie, Doug Oard and David Doermann Cross Language Entity Linking, in IJCNLP: International Joint Conference on Natural Language Processing. 2011. find
  2. Amor-Tijani, G., Enhanced english-arabic cross-language information retrieval. 2008, George Washington University find
  3. Winkler, W.E., Overview of record linkage and current research directions. 2006, BUREAU OF THE CENSUS. find
  4. Goiser, K. and P. Christen, Towards automated record linkage, in Proceedings of the fifth Australasian conference on Data mining and analystics - Volume 61. 2006, Australian Computer Society, Inc.: Sydney, Australia. p. 23-31. find
  5. Köpcke, H. and E. Rahm, Frameworks for entity matching: A comparison. Data & Knowledge Engineering, 2010. 69(2): p. 197-210. find
  6. Fellegi, I.P.S., A. B., A Theory for Record Linkage. Journal of the American Statistical Association, 1969. 64: p. 1183–1210. find
  7. Mohamed, K.A. and A. Hassan, Web usage mining analysis of federated search tools for Egyptian scholars. Program: electronic library and information systems, 2008. 42(4): p. 418-435. find
  8. ElYamany, H.F. and A.H. Yousef. A Mobile-Quiz Application in Egypt. in The 4th IEEE International E Learning Conference, Bahrain, 7-9 May, 2013. 2013. find
  9. El-Hadidi, M., et al. Quantifying the ICT Needs of Academic Institutes Using the Service Category-Stakeholder Matrix Approach. in Information & Communications Technology, 2008. ICICT 2008. ITI 6th International Conference on. 2008: IEEE. find
  10. Hussein, A.S., et al. e-Learning in the Egyptian Public Universities:Overview and Future Prospective. in ICT-Learn 2009 Conference, Human and Technology Development Foundation. 2009. find
  11. Christen, P. and K. Goiser, Quality and Complexity Measures for Data Linkage and Deduplication, in Quality Measures in Data Mining, F. Guillet and H. Hamilton, Editors. 2007, Springer Berlin / Heidelberg. p. 127-151. find
  12. Shaalan, K. and H. Raza, Person name entity recognition for Arabic, in Proceedings of the 2007 Workshop on Computational Approaches to Semitic Languages: Common Issues and Resources. 2007, Association for Computational Linguistics: Prague, Czech Republic. p. 17-24. find
  13. R.C.Russell, Russell Index U.S. Patent 1,261,167, http://patft.uspto.gov/netahtml/srchnum.htm. 1918. ↗
  14. R.C.Russell, Russell Index U.S. Patent 1,435,663, http://patft.uspto.gov/netahtml/srchnum.htm. 1922. ↗
  15. Aqeel, S.B., S.; Jensen, E.; Grossman, D.; Frieder, O., On the Development of Name Search Techniques for Arabic. Journal of the American Society of Information Science and Technology, 2006. 57(6). find
  16. Koujan, T. Arabic Soundex. 2008; Available from: http://www.codeproject.com/Articles/26880/Arabic-Soundex. ↗
  17. Levenshtein, V.I., Binary Codes Capable of Correcting Deletions, Insertions and Reversals. Soviet Physics Doklady, 1966. 10: p. 707-710. find
  18. Alvaro Monge and C. Elkan, The field matching problem: Algorithms and applications, in Second International Conference on Knowledge Discovery and Data Mining. 1996. p. 267-270. find
  19. Herzog, T.N., et al., Record Linkage – Methodology. 2007, Springer New York. p. 81-92. find
  20. Boussy, C.A., A comparison of hand and computer-linked records. 1992, University of Miami. find
  21. Yakout, M., M.J. Atallah, and A. Elmagarmid. Efficient Private Record Linkage. in Data Engineering, 2009. ICDE '09. IEEE 25th International Conference on. 2009. find
  22. Srinivasan, H., Machine learning for person identification with applications in forensic document analysis. 2008, State University of New York at Buffalo p. 182. find
  23. Dan Wu, D.H., Exploring the further integration of machine translation in English-Chinese cross language information access. Program: electronic library and information systems, 2012. 46(4): p. 429 - 457. find
  24. Jiampojamarn, S., Grapheme-to-phoneme conversion and its application to transliteration, in Department of Computing Science. 2010, University of Alberta. find
  25. Kim, H.-s., High Performance Record Linking, in Computer Science and Engineering. 2010, The Pennsylvania State University. find
  26. Christen, P., Automatic Training Example Selection for Scalable Unsupervised Record Linkage, in Proceedings of the 12th Pacific-Asia conference on Advances in knowledge discovery and data mining, PAKDD'08 T. Washio, et al., Editors. 2008, Springer Berlin / Heidelberg. p. 511-518. find
  27. Christen, P., Automatic record linkage using seeded nearest neighbour and support vector machine classification, in Proceeding of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining. 2008, ACM. p. 151-159. find
  28. Weifeng, S., W. Jiying, and F.H. Lochovsky, Record Matching over Query Results from Multiple Web Databases. Knowledge and Data Engineering, IEEE Transactions on, 2010. 22(4): p. 578-589. find
  29. Navarro, G., A guided tour to approximate string matching. ACM Computing Surveys (CSUR), 2001. 33(1): p. 31-88. find
  30. Al-onaizan, Y.K., Kevin, Machine Transliteration of Names in Arabic Text, in ACL Workshop on Comp. Approaches to Semitic Languages. 2002. p. 34-46. find
  31. Knight, K.G., Jonathan Machine Transliteration, in Computational Linguistics. 1997. p. 128-135. find
  32. AbdulJaleel, N.L., Leah S. , English to Arabic Transliteration for Information Retrieval: A Statistical Approach. 2003. find
  33. AbdulJaleel, N.L., Leah S. , Statistical transliteration for English-Arabic cross language information retrieval, in CIKM. 2003. p. 139-146. find
  34. Ma, X. and U.o. Pennsylvania, Improving Named Entity Recognition with Co-training and Unlabeled Bilingual Data. 2008: University of Pennsylvania. find
  35. Freeman, A.T., S.L. Condon, and C.M. Ackerman, Cross linguistic name matching in English and Arabic: a "one to many mapping" extension of the Levenshtein edit distance algorithm, in Proceedings of the main conference on Human Language Technology Conference of the North American Chapter of the Association of Computational Linguistics. 2006, Association for Computational Linguistics: New York, New York. p. 471-478. find
  36. Christen, P., Febrl: a freely available record linkage system with a graphical user interface, in Proceedings of the second Australasian workshop on Health data and knowledge management - Volume 80. 2008, Australian Computer Society, Inc.: Wollongong, NSW, Australia. p. 17-25. find
  37. Christen, P., Development and user experiences of an open source data cleaning, deduplication and record linkage system. SIGKDD Explor. Newsl., 2009. 11(1): p. 39-48. find
  38. Christen, P., T. Churches, and M. Hegland, Febrl – A Parallel Open Source Data Linkage System, in Advances in Knowledge Discovery and Data Mining, H. Dai, R. Srikant, and C. Zhang, Editors. 2004, Springer Berlin Heidelberg. p. 638-647. find
  39. Elfeky, M.G., V.S. Verykios, and A.K. Elmagarmid. TAILOR: a record linkage toolbox. in Data Engineering, 2002. Proceedings. 18th International Conference on. 2002. find
  40. Yancey, W.E., Bigmatch: A Program for Extracting Probable Matches from a Large File for Record Linkage, in Statistical Research Report Series RRC2002/01. 2002, US Bureau of the Census: Washington, D.C. find
  41. Dey, D., V.S. Mookerjee, and L. Dengpan, Efficient Techniques for Online Record Linkage. Knowledge and Data Engineering, IEEE Transactions on, 2011. 23(3): p. 373-387. find
  42. El-Shishtawy, T., A Hybrid Algorithm for Matching Arabic Names. arXiv preprint arXiv:1309.5657, 2013. ↗
  43. Higazy, A., et al. Web-based Arabic/English duplicate record detection with nested blocking technique. in Computer Engineering & Systems (ICCES), 2013 8th International Conference on. 2013. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/yousef-2014-cross-language-personal-name-mapping/