Self‐training author name disambiguation for information scarce scenarios

Abstract

We present a novel 3-step self-training method for author name disambiguation-SAND (self-training associative name disambiguator)-which requires no manual labeling, no parameterization (in real-world scenarios) and is particularly suitable for the common situation in which only the most basic information about a citation record is available (i.e., author names, and work and venue titles). During the first step, real-world heuristics on coauthors are able to produce highly pure (although fragmented) clusters. The most representative of these clusters are then selected to serve as training data for the third supervised author assignment step. The third step exploits a state-of-the-art transductive disambiguation method capable of detecting unseen authors not included in any training example and incorporating reliable predictions to the training data. Experiments conducted with standard public collections, using the minimum set of attributes present in a citation, demonstrate that our proposed method outperforms all representative unsupervised author grouping disambiguation methods and is very competitive with fully supervised author assignment methods. Thus, different from other bootstrapping methods that explore privileged, hard to obtain information such as self-citations and personal information, our proposed method produces topnotch performance with no (manual) training data or parameterization and in the presence of scarce information.

Combining Classifiers and User Feedback for Disambiguating Author Names

“…Step 1 aims at producing pure clusters for the next steps using the author and coauthor attributes, as in [1,5]. We use coauthorship relations among citations in order to group together those belonging to a same author.…”

A Graph Combination With Edge Pruning‐Based Approach for Author Name Disambiguation

“…Several existing works (Cota et al, ; Fan et al, ; Ferreira et al, ; Santana et al, ) report that overlapping coauthors between a pair of records with author name α is a strong indicator of author similarity. Thus, we initially generate an author‐similarity graph (a‐s graph) , formed by connecting node (article) pairs with overlapping coauthors.…”

“…A diverse range of techniques have been applied to the AND problem such as supervised approaches (support vector machines and naive Bayes: Han, Giles, Zha, Li, and Tsioutsiouliklis []), unsupervised approaches (Ferreira, Veloso, Gonçalves, & Laender, ; Khabsa, Treeratpituk, & Giles, ), graph‐based models (Markov random field: Tang, Fong, Wang, and Zhang []; factor graph model: Wang, Tang, Cheng, and Philip []), heuristic‐based solutions (Cota, Ferreira, Nascimento, Gonçalves, & Laender, ; Santana, Gonçalves, Laender, & Ferreira, ). Ferreira et al (2014), Liu, Li, Huang, and Fang (), and Cota et al (2010) have proposed grouping/clustering the records using coauthors, title, and venue.…”

Automatic Methods for Disambiguating Author Names in Bibliographic Data Repositories

“…According to our taxonomy, the methods may be classified following the main type of exploited approach: author grouping [3,9,11,12,14,13,16,17], which tries to group the references to the same author using some type of similarity among reference attributes, or author assignment [1,4,8,10,15,18], which aims at directly assigning the references to their respective authors. Alternatively, the methods may be grouped according to the evidence explored in the disambiguation task: the citation attributes (only), web information, or implicit data that can be extracted from the available information.…”

“…HHC disambiguates a set of citation records by successively fusing clusters of citation records with similar author names based on a real-world heuristic applied to their citation attributes. Then, we present SAND -Self-training Associative Name Disambiguator [9,8]. SAND is a three-step selftraining method for author name disambiguation that requires no manual labeling and no parameterization (in real world scenarios).…”