|
|
|
|
The Tenth q-bio Summer School - Albuquerque: Lecture 18
From Q-bio
Clustering Algorithms
Michael Wester (wester@math.unm.edu)
SpatioTemporal Modeling Center University of New Mexico Health Sciences Center
Clustering, PairCorr, SimulateDomains, SRcluster are a set of MATLAB classes that
- o collect together various spatial clustering statistics and algorithms
- (hierarchal, DBSCAN [4 versions], Getis based, Voronoi based);
- o pair auto- and cross-correlation curves and statistics
- o simulations of spatial domains (clusters) of fluorophore localizations that
- exhibit distributions of observations representing blinks; and
- o a top-down clustering algorithm to collapse clusters of observations of
- blinking fluorophores into a single estimate of the true location of the
- fluorophore using a log-likelihood hypothesis test (H-SET: Hierarchical
- Single Emitter Hypothesis Test).
Many of these algorithms operate in 3D as well as 2D. I will also demonstrate
SuperCluster, which provides a GUI interface to some of the above.
Please download MATLAB Clustering Classes Version 1, SuperCluster (and the DERIVEST Suite) from the links on
[[ http://stmc.health.unm.edu/tools-and-data/]]
Files for the demonstration
