A pragmatic approach to hierarchical categorization of research expertise in the presence of scarce information

Identifying collaboration dynamics of bipartite author-topic networks with the influences of interest changes

“…As co-authorship evolves over time, tie formations and changes in researchers' interest show the dynamics of co-author networks in longitudinal analysis. The interest changes over time is also accounted for measuring publication productivity (Amjad et al 2018) or finding researchers with temporal rank differentiated by publication venues (Daud et al 2010) since some venues have higher ratings (Meho 2019). In preceding studies on collaboration dynamics overtime, network properties like centrality or clustering coefficient were used to explain tie formations of researcher cliques (Abbasi et al 2011;Hou et al 2013).…”

Evaluating the state-of-the-art in mapping research spaces: A Brazilian case study

“…The Lattes dataset has been recurrently used by several works in the literature with a variety of objectives in the context of Brazilian science, such as characterizing co-authorship networks [ 18 ], proposing new metrics to quantify the influence of researchers [ 19 ], assessing the impact of academic mobility on the quality of graduate programs [ 20 ], studying the profiles of top researchers and graduate programs in Computer Science [ 21 , 22 ], and identifying patterns in interdisciplinary collaborations and analyzing their evolution [ 23 ]. The dataset has also been used to demonstrate the effectiveness of the method proposed in [ 24 ] for automatic research expertise classification from the titles of academic works. Nevertheless, to the best of our knowledge, this is the first time this data is used to map the Brazilian research space and to understand how this space has changed over time.…”

Current research on theory and practice of digital libraries: best papers from TPDL 2017

“…De Siqueira et al [1] in "A pragmatic approach to hierarchical categorization of research expertise in the presence of scarce information" use supervised machine learning methods trained with manually labeled information to associate researchers and their expertise to a knowledge classification scheme. This classification is performed on the minimum information available, and it is possible to enhance organization processes in digital libraries and repositories, as well as in recommendation systems.…”