(→Professor Bouman's "Cluster" Software) |
(→Professor Bouman's "Cluster" Software) |
||
| Line 1: | Line 1: | ||
== Professor Bouman's "Cluster" Software == | == Professor Bouman's "Cluster" Software == | ||
| − | The "Cluster" software, developed by Purdue University's own Professor Charles Bouman, can be found, along with all supporting documentation, here: [http://cobweb.ecn.purdue.edu/~bouman/software/cluster/ Cluster Homepage]. The algorithm takes an iterative bottom-up (or agglomerative) approach to clustering. Different from many clustering algorithms, this one uses a so-called "Rissaren criterion" or "minimum description length" (MDL) as the best fit criterion. In short, MDL favors density estimates with parameters that may be encoded with very few bits. i.e. The simpler it is to represent the density parameters in some binary form, the better the estimate is. | + | The "Cluster" software, developed by Purdue University's own Professor Charles Bouman, can be found, along with all supporting documentation, here: [http://cobweb.ecn.purdue.edu/~bouman/software/cluster/ Cluster Homepage]. The algorithm takes an iterative bottom-up (or agglomerative) approach to clustering. Different from many clustering algorithms, this one uses a so-called "Rissaren criterion" or "minimum description length" (MDL) as the best fit criterion. In short, MDL favors density estimates with parameters that may be encoded (along with the data) with very few bits. i.e. The simpler it is to represent both the density parameters and data in some binary form, the better the estimate is considered. |
Below is the block diagram from the software manual (with misplaced blocks corrected) describing the function of the software. | Below is the block diagram from the software manual (with misplaced blocks corrected) describing the function of the software. | ||
[[Image:BoumanClusterBlockDiagram_Old Kiwi.jpg]] | [[Image:BoumanClusterBlockDiagram_Old Kiwi.jpg]] | ||
Revision as of 11:09, 17 April 2008
Professor Bouman's "Cluster" Software
The "Cluster" software, developed by Purdue University's own Professor Charles Bouman, can be found, along with all supporting documentation, here: Cluster Homepage. The algorithm takes an iterative bottom-up (or agglomerative) approach to clustering. Different from many clustering algorithms, this one uses a so-called "Rissaren criterion" or "minimum description length" (MDL) as the best fit criterion. In short, MDL favors density estimates with parameters that may be encoded (along with the data) with very few bits. i.e. The simpler it is to represent both the density parameters and data in some binary form, the better the estimate is considered.
Below is the block diagram from the software manual (with misplaced blocks corrected) describing the function of the software.
