diff options
author | vrouvrea <vrouvrea@636b058d-ea47-450e-bf9e-a15bfbe3eedb> | 2019-01-17 16:15:59 +0000 |
---|---|---|
committer | vrouvrea <vrouvrea@636b058d-ea47-450e-bf9e-a15bfbe3eedb> | 2019-01-17 16:15:59 +0000 |
commit | 6b70186caa620f9cd734903db978b2117509f8b1 (patch) | |
tree | cb2cd36bfbeb480a6e7d7580c48591aa4d9c5440 /src/Rips_complex/doc/Intro_rips_complex.h | |
parent | a3b3e527f75ea05a7840ecf73f613f56c3c52273 (diff) | |
parent | e105da964b96218459c8822816de36273f6cdf17 (diff) |
Merge last trunk modifications
git-svn-id: svn+ssh://scm.gforge.inria.fr/svnroot/gudhi/branches/cubical_complex_small_fix@4062 636b058d-ea47-450e-bf9e-a15bfbe3eedb
Former-commit-id: 495752e4da032e54a1a88adbcd030f526bc80a43
Diffstat (limited to 'src/Rips_complex/doc/Intro_rips_complex.h')
-rw-r--r-- | src/Rips_complex/doc/Intro_rips_complex.h | 7 |
1 files changed, 5 insertions, 2 deletions
diff --git a/src/Rips_complex/doc/Intro_rips_complex.h b/src/Rips_complex/doc/Intro_rips_complex.h index 712d3b6e..a2537036 100644 --- a/src/Rips_complex/doc/Intro_rips_complex.h +++ b/src/Rips_complex/doc/Intro_rips_complex.h @@ -39,7 +39,7 @@ namespace rips_complex { * <a target="_blank" href="https://en.wikipedia.org/wiki/Vietoris%E2%80%93Rips_complex">(Wikipedia)</a> * is an abstract simplicial complex * defined on a finite metric space, where each simplex corresponds to a subset - * of point whose diameter is smaller that some given threshold. + * of points whose diameter is smaller that some given threshold. * Varying the threshold, we can also see the Rips complex as a filtration of * the \f$(n-1)-\f$dimensional simplex, where the filtration value of each * simplex is the diameter of the corresponding subset of points. @@ -53,7 +53,10 @@ namespace rips_complex { * The number of simplices in the full Rips complex is exponential in the * number of vertices, it is thus usually restricted, by excluding all the * simplices with filtration value larger than some threshold, and keeping only - * the dim_max-skeleton. + * the dim_max-skeleton. It may also be a good idea to start by making the + * point set sparser, for instance with + * `Gudhi::subsampling::sparsify_point_set()`, since small clusters of points + * have a disproportionate cost without affecting the persistence diagram much. * * In order to build this complex, the algorithm first builds the graph. * The filtration value of each edge is computed from a user-given distance |