Next:
Introduction
Up:
Abstract page
Contents
Introduction
The Self-Organizing Map
Structure
Initilialization
Training
Mathematical treatment
Voronoi region:
Convergence:
Energy function:
Important properties
Quantization and projection:
Errors in data:
Visualization
Vector components:
Clusters:
SOM in the input space:
Input vectors on the SOM:
Knowledge Discovery With the SOM
Visualization
Map visualization:
Object visualization:
Map measures
Topology
Resolution
Combined quality measures
Similarity between maps
Clustering
Clustering the SOM
Distance matrix
Hierarchical SOMs
Other clustering methods
Classification
Probability density estimation
Modeling
Lookup models
Local models
A Data Mining Tool
Data sets
Component scaling:
Histogram equalization:
Filtering:
Grasping the meaning:
Map training
Map visualization
Data on map
Postprocessing
Futher development
Pulp and Paper Industry
Pulp and paper technology
Paper machines and pulp lines
Mill techonology
Geographical areas
Scandinavian paper industry
Data sets
Environmental, economical and technological aspects
Combined maps
Conclusions
References
About this document ...
Juha Vesanto
Tue May 27 12:40:37 EET DST 1997