The NETWORK Procedure

This example looks at the structural relationship of US patent citations by using a large data set that is maintained by the Stanford Network Analysis Project (SNAP) (Leskovec 2014). The citation graph includes over 16 million citations made to patents between 1975 and 1999.

The following statements construct the links data table mycas.Patents from a local copy of the raw patent citation data:

filename in 'cit-Patents.txt';
data mycas.Patents;
   infile in firstobs=5 dlm='09'X;
   input from to;
run;

The following statements find the connected components of the citation graph by using a distributed union-find algorithm. This algorithm takes advantage of all the machines in your configured session.

proc network
   links        = mycas.Patents
   outNodes     = mycas.NodeSetOut;
   connectedComponents
      out       = mycas.ConCompOut
      algorithm = parallel;
run;
%put &_NETWORK_;

The progress of the procedure is shown in Output 2.12.1.

The 10 biggest components are shown in Output 2.12.2. It is interesting to note that the vast majority of patents (over 99%) are all contained in the same component. This is not too surprising, because many of the seminal patent claims are required in order to understand subsequent inventions.