Page 336 - Becoming Metric Wise
P. 336

328   Becoming Metric-Wise


          10.2.5 Reciprocity
          In a directed unweighted network, reciprocity is defined as the observed
          probability that if node i points to node j, also j points also to node i.A
          friendship network is a typical case for which it makes sense to calculate a
          reciprocity value.
             In weighted networks such as journal citation networks reciprocity is
          defined as follows (Squartini et al., 2013):
             Let w ij be the weight of the link between node i and node j, in that
          order, where w ii 5 0 for all i. Let rw ij 5 rw ji 5 min(w ij ,w ji ) denote the
          reciprocated weight between nodes i and j. Then the reciprocity value W
          for the network is:

                                         P P
                                           i  j  rw ij
                                                                     (10.12)
                                   W 5 P P
                                           i  j  w ij
          10.2.6 Complex Networks and the Small-World Phenomenon

          The term complex networks is not precisely defined, but usually refers to
          large networks which are neither regular nor completely random. One
          may say that the citation network is a complex network. Two well-known
          and much studied classes of complex networks are scale-free networks and
          small-world networks. The term scale-free network is used for networks
          that have a degree distribution following a power law. Size and rank fre-
          quency of hyperlinks between websites follow—approximately a power law
          as was observed by Rousseau (1997b) for a small sample and by Faloutsos
          et al. (1999) for a much larger set. Also the network of protein interactions
          has a power law distribution (Jeong et al., 2001). A small-world network is
          characterized as a network exhibiting a high degree of clustering and hav-
          ing at the same time a small average distance between nodes. These charac-
          teristics distinguish small-world networks from random networks which
          also have a small average distance between nodes, but a much smaller
          degree of clustering. The small-world phenomenon is related to the notion
          of six degrees of separation (Kochen, 1989; Milgram, 1967; Newman &
          Watts, 1999). The concept that all persons on earth are linked through
          friendship or acquaintanceship in at most six steps was first proposed in
          1929 by the Hungarian writer Frigyes Karinthy in a short story called
          “Chains” (Braun, 2004). This idea is a practical expression of the concept
          that such networks tend to grow exponentially. A related concept is the
          Erd˝ os number. The Erd˝ os number, named after the famous Hungarian
   331   332   333   334   335   336   337   338   339   340   341