Information Technology Reference
In-Depth Information
7.6
Summary
Typical citation-based domain visualization approaches have focused on citation
frequencies of high-profiled research in a knowledge domain. Consequently, resul-
tant visualizations are strongly biased towards highly cited works. Although highly
cited works constitute the core knowledge of a domain, its presence inevitably
outshines the presence of latent domain knowledge if we measure them with the
same yardstick. The use of two-step citation chains allows us to glean latent
domain knowledge and maintain the global picture of where such latent domain
knowledge fits.
In order to track the development of scientific paradigms, it is necessary to take
into account latent as well as mainstream domain knowledge. By incorporating
an information visualization procedure originally developed for visualizing main-
stream domain knowledge into a recursive process, it is possible for us to visualizing
not only highly relevant and highly cited documents, but also highly relevant but
infrequently cited documents.
A natural extension of the research is to explore ways that can combine
approaches based on citation patterns and those based on word-occurrence patterns
to pin point a significant mismatch between the citation strength and word
co-occurrence patterns. There are other potentially useful ways to uncover latent
domain knowledge. Many techniques developed in scientometrics for quantitative
studies of science can be used to generate structural representations of domain
knowledge. By comparing and contrasting differences across a variety of structural
representations one can expect to spot missing links and potentially noteworthy
connections. For example, if a co-word analysis reveals a strong link between
intellectually related works. In contrast, if such links are absent or weak in citation
networks, then it could be important for scientists to know whether they might have
overlooked something potentially significant.
On the one hand, visualizing domain knowledge in general is a revival of a
long established quest for quantitative studies of scientific discoveries and scientific
paradigms, especially due to the advances in enabling techniques such as digital
libraries and information visualization. On the other hand, visualizing domain
knowledge should set its own research agenda in the new era of science and
technology so as to provide valuable devices for scientists, philosophers of science,
sociologists of knowledge, librarians, government agencies, and others to grasp
crucial developments in science and technology.
In this chapter, we have examined the role of citation chains in visualizing
latent domain knowledge. The new visualization approach can not only capture the
intellectual structure of highly cited works but also make it possible to uncover con-
nections between latent domain knowledge and the body of the mainstream domain
knowledge. The two case studies have shown that this approach has the potential as
a new way of supporting knowledge tracking and knowledge management.
Search WWH ::




Custom Search