$k$-means considered harmful: On arbitrary topological changes in Mapper complexes
By: Mikael Vejdemo-Johansson
Potential Business Impact:
Makes data maps more honest and accurate.
The Mapper construction is one of the most widespread tools from Topological Data Analysis. There is an unfortunate trend as the construction has gained traction to use clustering methods with properties that end up distorting any analysis results from the construction. In this paper we will see a few ways in which widespread choices of clustering algorithms have arbitrarily large distortions of the features visible in the final Mapper complex.
Similar Papers
Multiscale 2-Mapper -- Exploratory Data Analysis Guided by the First Betti Number
Computational Geometry
Finds hidden patterns in complicated information.
When Annotators Disagree, Topology Explains: Mapper, a Topological Tool for Exploring Text Embedding Geometry and Ambiguity
Computation and Language
Shows how computers understand tricky words.
Tracking the Spatiotemporal Spread of the Ohio Overdose Epidemic with Topological Data Analysis
Applications
Finds drug overdose patterns to help stop them.