cluster grouping

Examples of cluster grouping in the following topics:

• Race and Genetics

  • Racial groups are sociologically, rather than biologically, different; that is to say, there is no "race" gene or set of genes.
  • Often, due to practices of group endogamy, allele frequencies cluster locally around kin groups and lineages, or by national, cultural, or linguistic boundaries - giving a detailed degree of correlation between genetic clusters and population groups when considering many alleles simultaneously.
  • While a person's race can generally be visually determined, different racial groups do not in fact differ biologically in substantial ways.

• Categorical REGE for directed binary data (Wasserman-Faust directed data)

  • A hierarchical clustering diagram can be useful if the equivalences found are inexact, or numerous, and a further simplification is needed.
  • Here, we see at level 2 of the clustering that there are three groups {A}, {B, C, D}, and {E, F, G, H, I}.
  • Should we want to use only two, however, the dendogram suggests that grouping A with B, C, and D would be the most reasonable choice.

• Equivalence of distances: Maxsim

  • Network>Roles & Positions>Automorphic>MaxSim generates a matrix of "similarity" between shape of the distributions of ties of actors that can be grouped by clustering and scaling into approximate classes.
  • Again, dimensional scaling or clustering of the distances can be used to identify sets of approximately automorphically equivalent actors.
  • The Euclidean distances between these lists are then created as a measure of the non-automorphic-equivalence, and hierarchical clustering is applied.
  • This small part of a large piece of output (there are 100 donors in the network) shows that a number of non-Indian casinos and race-tracks cluster together, and separately from some other donors who are primarily concerned with education and ecological issues.
  • The identification of approximate equivalence classes in valued data can be helpful in locating groups of actors who have a similar location in the structure of the graph as a whole.

• Clustering tools

  • Tools>Cluster>Hierarchical proceeds by initially placing each case in its own cluster.
  • This results in clusters of increasing size that always enclose smaller clusters.
  • "farthest neighbor") computes similarities between the member of the new cluster that is least similar to each other case not in the cluster.
  • This gives a clear picture of the similarity of cases, and the groupings or classes of cases.
  • The E-I index is often most helpful, as it measures the ratio of the numbers of ties within the clusters to ties between clusters.

• Clustering similarities or distances profiles

  • Cluster analysis is a natural method for exploring structural equivalence.
  • Tools>Cluster).
  • The second panel shows a rough character-mapped graphic of the clustering.
  • The dendogram can be particularly helpful in locating groupings of cases that are sufficiently equivalent to be treated as classes.
  • The measures of clustering adequacy in Tools>Cluster can provide additional guidance.

• Social Construct or Biological Lineage?

  • Both schemes benefitted the third group, the racially pure whites.
  • Specifically, their study utilized a software program that requires researchers to first decide how many clusters or groups they want the program to produce before it can analyze the data.
  • Other researchers, using the same data, found a different number of clusters from the same genetic data.
  • Indeed, the first medication marketed for a specific racial group, BiDil was recently approved by the U.S.
  • However, distinctions between racial groups are declining due to intermarriage and have been for years.

• Highlighting parts of the network

  • NetDraw graphs these sub-structures, and saves the information in the node-attribute database.Analysis>K-cores locates parts of the graph that form sub-groups such that each member of a sub-group is connected to N-K of the other members.
  • The graph is colored to represent the clusters, and database information is stored about the cluster memberships at various levels of aggregation.
  • A hierarchical clustering can be very interesting in understanding which groups are more homogeneous (those that group together at early stages in the clustering) than others; moving up the clustering tree diagram, we can see a sort of a "contour map" of the similarity of nodes.Analysis>Subgroups>Factions (select number).
  • The algorithm then forms the number of groups that you desire by seeking to maximize connection within, and minimize connection between the groups.
  • This is another numerical algorithm that seeks to create clusters of nodes that are closely connected within, and less connected between clusters.

• Value Clusters

  • People from different backgrounds tend to have different value systems, which cluster together into a more or less consistent system.
  • Certain values may cluster together into a more or less consistent system.
  • A communal or cultural value system is held by and applied to a community, group, or society.
  • Some sociologists are interested in better defining and measuring value clusters in different countries.
  • Their responses are aggregated and can be used to reveal regional value clusters, like those displayed in this map.

• Introduction

  • The whole idea of "equivalence" that we discussed in the last chapter is an effort to understand the pattern of relationships in a graph by creating classes, or groups of actors who are "equivalent" in one sense or another.
  • All of the methods for identifying such groupings are based on first measuring the similarity or dissimilarity of actors, and then searching for patterns and simplifications.
  • Multi-dimensional scaling and hierarchical cluster analysis are widely used tools for both network and non-network data.
  • That is, methods for identifying groups of nodes that are similar in their patterns of ties to all other nodes.
  • These methods (and those for other kinds of "equivalence" in the next two chapters) use the ideas of similarity/distance between actors as their starting point; and, these methods most often use clustering and scaling as a way of visualizing results.

• Two-mode correspondence analysis

  • Here, however, this difference means that the two groupings tend to participate in different groups of initiatives, rather than confronting one another in the same campaigns.
  • The lower right quadrant here contains a meaningful cluster of actors and events, and illustrates how the results of correspondence analysis can be interpreted.
  • The result is showing that there is a cluster of issues that "co-occur" with a cluster of donors - actors defining events, and events defining actors.