Next: Introduction
Up: Relationship-based Clustering and Cluster
Previous: List of Tables
  Contents
- All possible clusterings of up to objects into up to groups.
- Object-centered versus relationship-based clustering.
- Abstract overview of the general relationship-based,
single-layer, single-learner, batch clustering process.
- Example for the problem of scale.
- Illustration of the curse of dimensionality.
- The relationship-based clustering framework.
- Illustrative CLUSION patterns.
- Comparison of cluster visualization techniques.
- Visualizing partitioning drugstore customers from RETAIL data-set 1/2.
- Visualizing partitioning drugstore customers from RETAIL data-set 2/2.
- Comparison of various number of clusters for YAHOO news data:.
- Web-log session clustering using a vector space model
and cosine similarity versus using weighted longest common
sub-sequence similarity.
- Effect of sub-sampling on OPOSSUM.
- Properties of Euclidean-based, cosine, and extended
Jaccard similarity measures illustrated in 2 dimensions 1/2.
- Properties of Euclidean-based, cosine, and extended
Jaccard similarity measures illustrated in 2 dimensions 2/2.
- Confusion matrices illustrating quality
differences of RND, KM E, KM C, and GP C approaches on a sample of 800
documents from N20.
- Mutual information performance curves
comparing 10 algorithms on 4 data-sets.
- Amount of balancing achieved
for 4 data-sets in combination with 10 algorithms.
- Comparison of cluster quality in terms of
mutual information and balance on average over 4 data-sets
with 10 trials each at 800 samples.
- The cluster ensemble.
- Illustration of multiple views generating different clusterings.
- Illustration of Cluster-based Similarity Partitioning Algorithm
(CSPA).
- Illustration of HyperGraph Partitioning Algorithm
(HGPA).
- Illustration of Meta-CLustering Algorithm (MCLA).
- Comparison of consensus functions
in a controlled noise experiment.
- Detailed Robust Consensus Clustering (RCC) results.
- Summary of RCC results.
- Illustration of Feature-Distributed Clustering (FDC) on 8D5K
data 1/2.
- Illustration of Feature-Distributed Clustering (FDC) on 8D5K
data 2/2.
- Object-Distributed Clustering (ODC) results.
- Summary of contributions in this dissertation.
- 2D2K data-set.
- 8D5K data-set.
- IRIS data-set.
- Pen digits data-set PENDIG.
- PENDIG clusters.
- Drugstore data before preprocessing.
- Drugstore data sample RETAIL.
- Exemplary news web-page from YAHOO data-set.
- Exemplary newsgroup posting from N20 data-set.
Alexander Strehl
2002-05-03