1. Use GraphMatrixOperations in JUNG to get adjacency matrix M.
2. Use cern.colt.matrix.linalg.Algebra to compute an inverse matrix and
calculate (I-b*M)^{-1}-I, where I - an identity matrix, and b = 0.005.
3. i,j elements in the calculated matrix would give necessary coefficient
values.
4. Add a method to GraphMetricUtils and enhance GraphTopologyAnalyzerTest with
the appropriate test part.
See a paper "The Link Prediction Problem for Social Networks" for more details.
Original issue reported on code.google.com by nikita.z...@gmail.com on 3 May 2012 at 11:03
Original issue reported on code.google.com by
nikita.z...@gmail.com
on 3 May 2012 at 11:03