Mapping Science - Mapping Scientific Frontiers: The Quest for Knowledge Visualization - page 260

here: ones using a moving window of a fixed size [ Y - k , Y - 1 ] and ones using the
entire history ( Y o , Y - 1 ], where Y o is the earliest year of publication for records in
the given dataset.

8.1.3 Structural Variation Metrics

We expect that the degree of structural variation introduced by a new article can
offer prospective information because of the boundary spanning mechanism. If an
article introduces novel links that span the boundaries of different topics, then we
expect this signifies its potential in taking the intellectual structure for a new turn.
Given a baseline network, structural variations can be measured based on
information provided by a particular article. We will introduce three metrics of
structural variation. Each metric quantifies the degree of change in the baseline
network introduced by information provided by an article. No usage data is involved
in the measurement. The three metrics are modularity change rate, inter-cluster
linkage, and centrality divergence. The definitions of the first two metrics depend
on a partition of the baseline network, but the third one does not. A partition
of a network decomposes the network into non-overlapping groups of nodes. For
example, clustering algorithms such as spectral clustering can be used to partition a
network.
The theoretical underpinning of the structural variation is that scientific discov-
eries, at least a subset of them, can be explained in terms of boundary spanning,
brokerage, and synthesis mechanisms in an intellectual space (Chen et al. 2009 ).
This conceptualization generalizes the principle of literature-based discovery pio-
neered by Swanson ( 1986a , b ), which assumes that connections between previously
disparate bodies of knowledge are potentially valuable. In Swanson's famous ABC
model, the relationships AB and BC are known in the literature. The potential
relationship AC becomes a candidate that is subject to further scientific investigation
(Weeber 2003 ). Our conceptualization is more generic in several ways. First,
in the ABC model, the AC relation changes an indirect connection to a direct
connection, whereas our structural variation model makes no assumption about
any prior relations at all. Second, in the ABC model, the scope of consideration is
limited to relationships involving three entities. In contrast, our structural variation
model takes a wider context into consideration and addresses the novelty of a
connection that links groups of entities as well as connections linking individual
entities. Because of the broadened scope of consideration, it becomes possible to
search for candidate connections more effectively. In other words, given a set of
entities, the size of the search space of potential connections can be substantially
reduced if additional constraints are applicable for the selection of candidate
connections. For example, the structural hole theory developed in social network
analysis emphasizes the special potential of nodes that are strategically positioned
to form brokerage, or boundary spanning, links and create good ideas (Burt 2004 ;
Chen et al. 2009 ).