Search Results

You are looking at 1 - 8 of 8 items for :

  • "Hybrid clustering" x
  • All content x
Clear All

and by which venues, extract “hot topics” and predict trends. Hybrid clustering refers to the clustering of the same class of entities with multi-view representations, either from various information sources or from different feature generators (we

Restricted access

window, where, of course, the same (hybrid) clustering algorithm should be applied. After the clustering has been conducted in each time window separately, all obtained clusters or classes representing research topics in the corresponding time slice form

Restricted access

to apply core documents to represent and to describe publication clusters, first a hybrid cluster analysis of two selected ISI Subject Categories has been conducted. In a second step, an analysis of core documents in the corresponding ISI Subject

Restricted access

application of the degree h-index to topic representation. The hybrid clustering (cf. Glänzel and Thijs 2011b) of the subject category “geography” for the period 2004–2008 provided a topic, that has been labelled urban political ecology . The core documents

Restricted access

, creative and business. Consequently, the combination of textual data and citation data is thought as a promising method to deal with scientific publication. Some hybrid clustering has been carried out, such as Janssens et al. ( 2006a , b ) put forward a

Restricted access
Authors: Frizo Janssens, Wolfgang Glänzel, and Bart De Moor


Previous studies have shown that hybrid clustering methods that incorporate textual content and bibliometric information can outperform clustering methods that use only one of these components. In this paper we apply a hybrid clustering method based on Fisher’s inverse chisquare to integrate full-text with citations and to provide a mapping of the field of information science. We quantitatively and qualitatively asses the added value of such an integrated analysis and we investigate whether the clustering outcome is a better representation of the field by comparing with a text-only clustering and with another hybrid method based on linear combination of distance matrices. Our data set consists of almost 1000 articles and notes published in the period 2002–2004 in 5 representative journals. The optimal number of clusters for the field is 5, determined by using a combination of distance-based and stability-based methods. Term networks present the cognitive structure of the field and are complemented by the most representative publications. Three large traditional sub-disciplines, particularly, information retrieval, bibliometrics/scientometrics, and more social aspects, and two smaller clusters about patent analysis and webometrics, can be distinguished.

Restricted access

upper crust using joint orientation distributions . Journal of Structural Geology , 27 ( 10 ), 1778 – 1787 . 10.1016/j.jsg.2005.05.016 Yaghini , M. , Soltanian , R. , & Noori , J. ( 2012 ). A hybrid clustering method using genetic algorithm

Open access

. Janssens , F. , Glänzel , W. , de Moor , B. ( 2007 ). Dynamic hybrid clustering of bioinformatics by incorporating text mining and citation analysis . In P. Berkhin , R. Caruana , X. Wu , S. Gaffney (eds.), Proceedings of the 13th ACM SIGKDD

Restricted access