Inverse_X issue and cluster select.

[WolVesz]

Hello All,

Really like this project, better than all the other mapper tools on python right now imo.

However, I am consistently having a problem where whenever I include inverse_X in the mapping, the mapping returns 0 nodes and 0 edges. I have tried this with a number of different data sets without success. I have done this on data sets with 300D and 200,000 rows and 200D, 100D, and all your sample sets with the same result. Its possible I am just ignorant, but I believe that all of these data sets should work with Inverse_X given it is just projecting along the original data. There are no errors produced for this problem. If inverse_X is not included, the mapper works as expected.

On another note, I am going to start working on a right click selection tool for displaying contents of nodes as i need this for my research. No idea how successful i'll be. Ill let you know.

[sauln]

Hi!

Thanks for the feedback. I look forward to seeing how the right click tool works out.

I'm not sure what you mean by include inverse_X in the mapping? Are you including it as the second argument of the map method? Or by not including a second argument? If you want to build mapper on your data without using a lens or filter function, it should work just fine

Here is how you would do that with the cat example.

data = np.genfromtxt('cat-reference.csv',delimiter=',')
graph = mapper.map(data,
                   clusterer=sklearn.cluster.DBSCAN(eps=0.1, min_samples=5),
                   coverer=km.Cover(nr_cubes=15, overlap_perc=0.2))

Is this what you're looking for?

[WolVesz]

I am referring to the first case of the above photo. If i had just passed the projected_data, it would have worked just fine (shown in second photo). The inclusion of the Inverse_X function causes 0 edges and 0 nodes to be developed (shown in first photo).

[sauln]

I suspect the parameters for DBSCAN are too small when it's clustering in the full 161 dimensions. For example, if I lower eps in the cat example, I also get 0 nodes and 0 edges:

>>> graph = mapper.map(lens,
                   inverse_X=data,
                   clusterer=sklearn.cluster.DBSCAN(eps=0.0001, min_samples=5),
                   coverer=km.Cover(nr_cubes=15, overlap_perc=0.2))
Created 0 edges and 0 nodes in 0:00:00.052353.

[MLWave]

It can also be the min_samples = 5, because with 15.625 hypercubes it may be not any have 5 or more samples in them (you can see this by setting verbose=2).

If you want to force nodes and links, without worrying about eps, try clustering with sklearn.cluster.KMeans(2) or probably better sklearn.cluster.AgglomerativeClustering(2). hdbscan.HBSCAN() also does not require guessing the correct eps.

[sauln]

@WolVesz were you able to get this working?

[sauln]

I assume the issue has been cleared up. If not, please reopen and we will figure out what's wrong.