Scaling Author Name Disambiguation with CNF Blocking

An author name disambiguation (AND) algorithm identifies a unique author entity record from all similar or same publication records in scholarly or similar databases. Typically, a clustering method is used that requires calculation of similarities between each possible record pair. However, the total number of pairs grows quadratically with the size of the author database making such clustering difficult for millions of records. One remedy for this is a blocking function that reduces the number of pairwise similarity calculations. Here, we introduce a new way of learning blocking schemes by using a conjunctive normal form (CNF) in contrast to the disjunctive normal form (DNF). We demonstrate on PubMed author records that CNF blocking reduces more pairs while preserving high pairs completeness compared to the previous methods that use a DNF with the computation time significantly reduced. Thus, these concepts in scholarly data can be better represented with CNFs. Moreover, we also show how to ensure that the method produces disjoint blocks so that the rest of the AND algorithm can be easily paralleled. Our CNF blocking tested on the entire PubMed database of 80 million author mentions efficiently removes 82.17% of all author record pairs in 10 minutes.

[1]  Satoshi Sekine,et al.  A survey of named entity recognition and classification , 2007 .

[2]  Lynette Hirschman,et al.  Appendix F: MUC-7 Coreference Task Definition (version 3.0) , 1998, MUC.

[3]  Marcos André Gonçalves,et al.  A brief survey of automatic methods for author name disambiguation , 2012, SGMD.

[4]  Neil R. Smalheiser,et al.  Author name disambiguation in MEDLINE , 2009, TKDD.

[5]  Peter Christen,et al.  A Survey of Indexing Techniques for Scalable Record Linkage and Deduplication , 2012, IEEE Transactions on Knowledge and Data Engineering.

[6]  Yong Yu,et al.  Leveraging Unlabeled Data to Scale Blocking for Record Linkage , 2011, IJCAI.

[7]  Madian Khabsa,et al.  Random Forest DBSCAN Clustering for USPTO Inventor Name Disambiguation and Conflation , 2016 .

[8]  C. Lee Giles,et al.  Adaptive sorted neighborhood methods for efficient record linkage , 2007, JCDL '07.

[9]  Ivan P. Fellegi,et al.  A Theory for Record Linkage , 1969 .

[10]  Craig A. Knoblock,et al.  Learning Blocking Schemes for Record Linkage , 2006, AAAI.

[11]  Ashwin Machanavajjhala,et al.  An automatic blocking mechanism for large-scale de-duplication tasks , 2012, CIKM '12.

[12]  Qing Wang,et al.  A Clustering-Based Framework to Control Block Sizes for Entity Resolution , 2015, KDD.

[13]  Lifang Gu,et al.  Adaptive Filtering for Efficient Record Linkage , 2004, SDM.

[14]  Raymond J. Mooney,et al.  Adaptive Blocking: Learning to Scale Up Record Linkage , 2006, Sixth International Conference on Data Mining (ICDM'06).

[15]  Wanli Liu,et al.  Author Name Disambiguation for PubMed , 2013, J. Assoc. Inf. Sci. Technol..

[16]  Madian Khabsa,et al.  Large scale author name disambiguation in digital libraries , 2014, 2014 IEEE International Conference on Big Data (Big Data).

[17]  Salvatore J. Stolfo,et al.  The merge/purge problem for large databases , 1995, SIGMOD '95.

[18]  R. Mooney Encouraging Experimental Results on Learning CNF , 1995 .

[19]  Madian Khabsa,et al.  Online Person Name Disambiguation with Constraints , 2015, JCDL.

[20]  Andrew McCallum,et al.  Efficient clustering of high-dimensional data sets with application to reference matching , 2000, KDD '00.

[21]  Daniel P. Miranker,et al.  An Unsupervised Algorithm for Learning Blocking Schemes , 2013, 2013 IEEE 13th International Conference on Data Mining.

[22]  Kush R. Varshney,et al.  Learning sparse two-level boolean rules , 2016, 2016 IEEE 26th International Workshop on Machine Learning for Signal Processing (MLSP).

[23]  Keizo Oyama,et al.  A Fast Linkage Detection Scheme for Multi-Source Information Integration , 2005, International Workshop on Challenges in Web Information Retrieval and Integration.

[24]  Daniel Jurafsky,et al.  Citation-based bootstrapping for large-scale author disambiguation , 2012, J. Assoc. Inf. Sci. Technol..

[25]  Madian Khabsa,et al.  Inventor name disambiguation for a patent database using a random forest and DBSCAN , 2016, 2016 IEEE/ACM Joint Conference on Digital Libraries (JCDL).

[26]  Avigdor Gal,et al.  Comparative Analysis of Approximate Blocking Techniques for Entity Resolution , 2016, Proc. VLDB Endow..