The Intelius Nickname Collection: Quantitative Analyses from Billions of Public Records

2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 607–610, Montr´eal, Canada, June 3-8, 2012. c⃝2012 Association for Computational Linguistics The Intelius Nickname Collection: Quantitative Analyses from Billions of Public Records Vitor R. Carvalho, Yigit Kiran and Andrew Borthwick Intelius Data Research 500 108th Avenue NE, Bellevue, WA 98004 {vcarvalho,ykiran,aborthwick}@intelius.com Abstract Although first names and nicknames in the United States have been well documented, there has been almost no quantitative analysis on the usage and association of these names amongst themselves. In this paper we in- troduce the Intelius Nickname Collection, a quantitative compilation of millions of name- nickname associations based on information gathered from billions of public records. To the best of our knowledge, this is the largest collection of its kind, making it a natural re- source for tasks such as coreference resolu- tion, record linkage, named entity recogni- tion, people and expert search, information ex- traction, demographic and sociological stud- ies, etc. The collection will be made freely available. 1 Introduction Nicknames are descriptive, invented person names that are frequently used in addition or instead of the person’s official name. Very often nicknames are truncated forms of the original name that can be used for convenience — for instance, ’Betsy’ instead of ’Elizabeth’. Previous studies on nicknames have mostly fo- cused on their origins or common descriptions. The Oxford Dictionary of First Names (Hanks et al., 2007), for instance, presents a comprehensive de- scription of origins and common uses of most nick- names in modern English. More quantitative explo- rations of the subject, such as the one provided by Alias Conditional Probability Betty 4.51% Beth 3.83% Liz 3.34% Elisabeth 0.95% Betsy 0.92% Table 1: Nickname Distribution Sample for “Elizabeth” the US Social Security Office1 tend to focus on baby name selection and on the relative popularity of most common first names. In this paper we present a quantitative study on nickname usage in the United States. Using bil- lions of personal public records and a state-of-the- art large-scale record linkage system, we were able to generate a comprehensive dataset with millions of name-nickname associations and their relative strength. A small sample of this collection can be seen in Table 1, where the most frequent nick- names associated with the first name “Elizabeth” and their Conditional Alias Probabilities. We ex- plain the derivation of these probabilities in detail in Section 3.3. This collection can provide valu- able features and insights for applications as diverse as entity extraction, coreference resolution, people search, language modeling, and machine translation. It will be made freely available for download from the Linguistic Data Consortium. 1Popular Baby Names from Social Security Online: http://www.ssa.gov/OACT/babynames/ 607 2 Prior Work To the best of our knowledge, there are no com- prehensive, empirically derived nickname databases currently made freely available for research pur- poses. (Bollacker, 2008) contains an extensive database of names and nicknames2, with listings on over 13,000 given names, containing multi- ple “variations” for each name. However, this …

Content truncated to 500 words