Premium
Enhanced self‐citation detection by fuzzy author name matching and complementary error estimates
Author(s) -
Donner Paul
Publication year - 2016
Publication title -
journal of the association for information science and technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.903
H-Index - 145
eISSN - 2330-1643
pISSN - 2330-1635
DOI - 10.1002/asi.23399
Subject(s) - citation , computer science , benchmark (surveying) , string searching algorithm , matching (statistics) , fuzzy logic , information retrieval , data mining , string (physics) , recall , artificial intelligence , pattern matching , statistics , mathematics , world wide web , psychology , cognitive psychology , geodesy , mathematical physics , geography
In this article I investigate the shortcomings of exact string match‐based author self‐citation detection methods. The contributions of this study are twofold. First, I apply a fuzzy string matching algorithm for self‐citation detection and benchmark this approach and other common methods of exclusively author name‐based self‐citation detection against a manually curated ground truth sample. Near full recall can be achieved with the proposed method while incurring only negligible precision loss. Second, I report some important observations from the results about the extent of latent self‐citations and their characteristics and give an example of the effect of improved self‐citation detection on the document level self‐citation rate of real data.