Mapper Lifting with Connectivity Test (Pointcloud to Graph) #67
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The algorithm initially constructs the Mapper graph from the given point cloud. Each vertex$v$ in the graph is associated with a set of points $\phi(v)$ , and two vertices $(u, v)$ are connected if their point sets intersect. Our connectivity test determines whether there is significant evidence for the connectedness of $\phi(u)$ and $\phi(v)$ .
We formulate the connectivity test using a recently observed universal property of persistent diagrams [1], which enables us to detect statistically significant homological cycles. The test employs "Weak Universality" and calculates the number of significant relative cycles in$H_1(\phi(u) \cup \phi(v), \phi(u) \setminus \phi(v) \cup \phi(v) \setminus \phi(u))$ as well as the number of significant cycles in $H_1(\phi(u) \cap \phi(v))$ . The emergence of new relative cycles confirms the connectivity between $u$ and $v$ .
[1] Bobrowski, O., Skraba, P. A universal null-distribution for topological data analysis. Sci Rep 13, 12274 (2023).
Code for loading and generating point clouds is adapted from #34.