Combining domain-specific heuristics for author name disambiguation

Abstract

Author name disambiguation has been one of the hardest problems faced by digital libraries since their early days. Historically, supervised solutions have empirically outperformed those based on heuristics, but with the burden of having to rely on manually labelled training sets for the learning process. Moreover, most supervised solutions just apply some type of generic machine learning solution and do not exploit specific knowledge about the problem. In this paper, we follow a similar reasoning, but in the opposite direction. Instead of extending an existing supervised solution, we propose a set of carefully designed heuristics and similarity functions and apply supervision only to optimize such parameters for each particular dataset. As our experiments show, the result is a very effective, efficient and practical author name disambiguation method that can be used in many different scenarios.

Random Forest DBSCAN for USPTO Inventor Name Disambiguation

“…Fan et al [5] used graph based framework for name disambiguation, and Hermansson et al [10] used graph kernels to calculate similarity based on local neighborhood structure. Instead of using machine learning algorithms, Santana et al [18] used domain specific heuristics for classification. Recently, Ventura et al [23] applied a random forest classifier with agglomerative clustering for inventor name disambiguation for USPTO database.…”

Heurísticas para desambiguação incremental de nomes de autores em referências bibliográficas

“…Uma parte dos trabalhos desenvolvidos foi publicada e apresentada na ACM/IEEE Joint Conference on Digital Libraries (JCDL) [Santana et al 2014] (Qualis A2), a principal conferência mundial naárea. Uma versão estendida desse trabalho foi convidada para uma edição especial dos melhores artigos da conferência e aceita para publicação, após nova revisão, no periódico International Journal on Digital Libraries (IJDL, Qualis B3) [Santana et al 2015] .…”

Online Person Name Disambiguation with Constraints

“…[23] and [12] use heuristics to automatically generate reference sets and use them, instead of labeled examples, as training data for the supervised methods. In [20] disambiguation is conducted using heuristics, with supervision being applied to optimize the heuristics parameters. We can also classify person name disambiguation methods based on the types of information they employ.…”