Comments on: Algorithms for Matching Strings http://semantichumanities.wordpress.com/2006/08/30/algorithms-for-matching-name-variants/ web technology and humanities scholarship Fri, 23 Oct 2009 15:15:04 +0000 http://wordpress.com/ hourly 1 By: AAA http://semantichumanities.wordpress.com/2006/08/30/algorithms-for-matching-name-variants/#comment-120 AAA Sat, 25 Nov 2006 02:14:24 +0000 https://semantichumanities.wordpress.com/2006/08/30/algorithms-for-matching-name-variants/#comment-120 It would be great if you publish all your refs. Thanks. It would be great if you publish all your refs. Thanks.

]]> By: Gene T http://semantichumanities.wordpress.com/2006/08/30/algorithms-for-matching-name-variants/#comment-67 Gene T Mon, 04 Sep 2006 13:34:13 +0000 https://semantichumanities.wordpress.com/2006/08/30/algorithms-for-matching-name-variants/#comment-67 There's a lot of string/sequence matching algo's out there: http://www.dcs.shef.ac.uk/~sam/stringmetrics.html http://www-igm.univ-mlv.fr/~lecroq/string/index.html (i can send you dozens mroe refs is you're interested) http://homepages.cs.ncl.ac.uk/brian.randell/Genealogy/NameMatching.pdf If you want to dig into Computational Linguistics (Info Extraction and Named ENtity Recognition), the state of the art are techniques like max entropy, conditional random fields, e.g. http://www.cs.umass.edu/~mccallum/papers/collseg04icmlws.pdf There’s a lot of string/sequence matching algo’s out there:

http://www.dcs.shef.ac.uk/~sam/stringmetrics.html
http://www-igm.univ-mlv.fr/~lecroq/string/index.html
(i can send you dozens mroe refs is you’re interested)
http://homepages.cs.ncl.ac.uk/brian.randell/Genealogy/NameMatching.pdf

If you want to dig into Computational Linguistics (Info Extraction and Named ENtity Recognition), the state of the art are techniques like max entropy, conditional random fields, e.g.

http://www.cs.umass.edu/~mccallum/papers/collseg04icmlws.pdf

]]> By: Got http://semantichumanities.wordpress.com/2006/08/30/algorithms-for-matching-name-variants/#comment-62 Got Fri, 01 Sep 2006 11:21:54 +0000 https://semantichumanities.wordpress.com/2006/08/30/algorithms-for-matching-name-variants/#comment-62 This kind of research is very interesting. I am trying it in order to request corpus in medieval latin or in ancient french and it can represent a solution if the corpus are not lemmatized. Levenstein distance, in particular, is the base algorithm for fuzzy searche and is implemented in exist, an XML database : http://exist-db.org/ (and the fuzzy search : http://demo.exist-db.org/xquery/functions.xq#text:fuzzy-index-terms). Maybe this implementation can give you some idea. This kind of research is very interesting. I am trying it in order to request corpus in medieval latin or in ancient french and it can represent a solution if the corpus are not lemmatized. Levenstein distance, in particular, is the base algorithm for fuzzy searche and is implemented in exist, an XML database : http://exist-db.org/ (and the fuzzy search : http://demo.exist-db.org/xquery/functions.xq#text:fuzzy-index-terms). Maybe this implementation can give you some idea.

]]> By: semantichumanities http://semantichumanities.wordpress.com/2006/08/30/algorithms-for-matching-name-variants/#comment-61 semantichumanities Thu, 31 Aug 2006 08:47:37 +0000 https://semantichumanities.wordpress.com/2006/08/30/algorithms-for-matching-name-variants/#comment-61 Yes, I can imagine where co-authorship is quite common, like in scientific articles, you could do pretty interesting things with networks of co-authorship. Perhaps a handy OPAC feature would be an interface that visualised this and let you navigate through it (see http://www.openacademia.org ). Or a search that let you limit a search on one author by the nth degrees of separation from another. Yes, I can imagine where co-authorship is quite common, like in scientific articles, you could do pretty interesting things with networks of co-authorship. Perhaps a handy OPAC feature would be an interface that visualised this and let you navigate through it (see http://www.openacademia.org ). Or a search that let you limit a search on one author by the nth degrees of separation from another.

]]> By: Pierre http://semantichumanities.wordpress.com/2006/08/30/algorithms-for-matching-name-variants/#comment-60 Pierre Wed, 30 Aug 2006 19:14:19 +0000 https://semantichumanities.wordpress.com/2006/08/30/algorithms-for-matching-name-variants/#comment-60 Resolving names can be an headache, for example a author named "Chen C" matches more than <a href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=pubmed&cmd=Search&itool=pubmed_AbstractPlus&term=%22Chen+C%22%5BAuthor%5D" rel="nofollow">2400 references</a> on <a href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed" rel="nofollow">pubmed</a>. The coauthorship could be used to discriminate one author from another ? Pierre Resolving names can be an headache, for example a author named “Chen C” matches more than 2400 references on pubmed.

The coauthorship could be used to discriminate one author from another ?

Pierre

]]>