Computer Science > Machine Learning
[Submitted on 4 Jun 2020 (v1), revised 11 Jun 2020 (this version, v3), latest version 15 Feb 2021 (v5)]
Title:Fuzzy c-Means Clustering for Persistence Diagrams
View PDFAbstract:Persistence diagrams, a key tool in the field of Topological Data Analysis, concisely represent the topology of a point cloud. Most current methods to integrate persistence diagrams into machine learning either require prior knowledge of a ground-truth topology or map them into a feature vector, offering a trade-off between information loss and invoking the curse of dimensionality. In this paper we give an algorithm for Fuzzy c-Means (FCM) clustering directly on the space of persistence diagrams, enabling unsupervised learning that automatically captures the topological structure of data, with no prior knowledge or additional processing of persistence diagrams. We prove the same convergence guarantees as traditional FCM clustering: that any convergent subsequence of our algorithm tends to a local minimum or saddle point of the cost function. We end by presenting experiments that demonstrate our algorithm can successfully cluster transformed datasets from materials science where comparable Wasserstein barycentre clustering algorithms fail, whilst also running at least an order of magnitude faster.
Submission history
From: Thomas Davies [view email][v1] Thu, 4 Jun 2020 11:45:20 UTC (1,993 KB)
[v2] Fri, 5 Jun 2020 13:26:52 UTC (1,992 KB)
[v3] Thu, 11 Jun 2020 12:15:03 UTC (1,993 KB)
[v4] Tue, 15 Dec 2020 21:19:26 UTC (1,393 KB)
[v5] Mon, 15 Feb 2021 13:00:53 UTC (4,386 KB)
Current browse context:
cs.LG
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
IArxiv Recommender
(What is IArxiv?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.