Advanced visualization of Self-Organizing Maps with vector fields

Pölzlbauer, Georg; Dittenbach, Michael; Rauber, Andreas

doi:10.1016/j.neunet.2006.05.013

Cited by 38 publications

(22 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Further analyses are now easier because the amount of data is reduced and patterns in the data are more evident. In an exploratory data-analysis, the interpretation of SOM is regularly based on illustrations, where quantitative information is depicted as a colour values on the map lattice (Pölzlbauer et al, 2006). Those maps, called component planes, reveal the distribution of each variable of original data on the SOM.…”

Section: > Statistical Analyses and Modellingmentioning

confidence: 99%

“…The SOM is an unsupervised artificial neural network especially feasible in an exploratory data analysis. The SOM can be used as a data visualization tool that performs a non-linear projection from a high-dimensional feature space onto a twodimensional map lattice (Pölzlbauer et al, 2006). In practice, large multivariable data can be presented to the user as intuitive pictures i.e.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Analysing a large dataset on long-term monitoring of water quality and plankton with the SOM clustering

Voutilainen

Rahkola-Sorsa

Parviainen

et al. 2012

Knowl. Managt. Aquatic Ecosyst.

View full text Add to dashboard Cite

Key-words:boreal lake, chlorophyll, phytoplankton, Self-Organizing map, water temperature, zooplanktonThe Self-Organizing Map (SOM) proved to be the method of choice for analysing a large heterogeneous ecological dataset. In addition to distributing the data into clusters, the SOM enabled hunting for correlations between the data components. This revealed logical and plausible relationships between and within the environment and groups of organisms. The main conclusions derived from the results were: (i) the structure of early summer plankton community significantly differed from that of late summer community in Lake Pyhäselkä and (ii) plankton community in late summer was characterized by two functional groups. The first group was formed mainly by phytoplankton, rotifers, and small cladocerans, such as Bosmina spp., and driven by water temperature. The second group was formed by small copepods and the abundant generalist herbivorous cladocerans Daphnia cristata and Limnosida frontosa, which, in turn, associated with chlorophyll a concentration. Biomasses of Bosmina spp. and D. cristata showed decreasing monotonic trends during a 20-year study period supposedly due to oligotrophication. Versatile possibilities to cluster data and hunt for correlations between data components offered by the SOM decisively helped to reveal associations across the original variables and draw conclusions. The results would have been undetectable solely on the basis of unorganised values. RÉSUMÉAnalyse d'un large ensemble de données de surveillance à long terme de la qualité de l'eau et du plancton par classification SOM Mots-clés :lac boréal, phytoplancton, carte auto-organisatrice, La carte d'auto-organisation (SOM) s'est avérée être la méthode de choix pour l'analyse d'un large ensemble de données hétérogènes écologiques. En plus de distribuer les données en grappes, la SOM permet la recherche de corrélations entre les composantes des données. Cette étude a révélé des relations logiques et plausibles entre et au sein de l'environnement et des groupes d'organismes. Les principales conclusions tirées des résultats étaient les suivantes : (i) la structure de la communauté planctonique du début de l'été diffère significativement de celle de la communauté de fin de l'été dans le lac Pyhäselkä et (ii) la communauté planctonique en fin d'été a été marquée par deux groupes fonctionnels. Le premier

show abstract

Section: > Statistical Analyses and Modellingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Analysing a large dataset on long-term monitoring of water quality and plankton with the SOM clustering

Voutilainen

Rahkola-Sorsa

Parviainen

et al. 2012

Knowl. Managt. Aquatic Ecosyst.

View full text Add to dashboard Cite

show abstract

“…Also, shape-based and vector-based SOM cluster visualizations have been introduced. 9,19 Further approaches using color to distinguish between clusters exist. 17 In addition, this cluster visualizations enable detecting topographic errors on the map, as SOM prototypes with similar vector attributes are colored with similar color, accordingly.…”

Section: Som Visualization and Applicationsmentioning

confidence: 99%

Multiscale visual quality assessment for cluster analysis with self-organizing maps

Bernard

Landesberger

Bremm

et al. 2011

Visualization and Data Analysis 2011

View full text Add to dashboard Cite

Cluster analysis is an important data mining technique for analyzing large amounts of data, reducing many objects to a limited number of clusters. Cluster visualization techniques aim at supporting the user in better understanding the characteristics and relationships among the found clusters. While promising approaches to visual cluster analysis already exist, these usually fall short of incorporating the quality of the obtained clustering results. However, due to the nature of the clustering process, quality plays an important aspect, as for most practical data sets, typically many different clusterings are possible. Being aware of clustering quality is important to judge the expressiveness of a given cluster visualization, or to adjust the clustering process with refined parameters, among others.In this work, we present an encompassing suite of visual tools for quality assessment of an important visual cluster algorithm, namely, the Self-Organizing Map (SOM) technique. We define, measure, and visualize the notion of SOM cluster quality along a hierarchy of cluster abstractions. The quality abstractions range from simple scalar-valued quality scores up to the structural comparison of a given SOM clustering with output of additional supportive clustering methods. The suite of methods allows the user to assess the SOM quality on the appropriate abstraction level, and arrive at improved clustering results. We implement our tools in an integrated system, apply it on experimental data sets, and show its applicability.

show abstract

“…The U-Matrix [7] shows local cluster boundaries by depicting pair-wise distances of neighbouring model vectors. The Gradient Field [4] has some similarity with the U-Matrix, but applies smoothing over a broader neighbourhood. It uses a vector field style of representation, where each arrow points to its closest cluster centre.…”

Section: Self-organising Map and Visualisationsmentioning

confidence: 99%

Visualising Clusters in Self-Organising Maps with Minimum Spanning Trees

Mayer

Rauber

2010

Artificial Neural Networks – ICANN 2010

Self Cite

View full text Add to dashboard Cite

Abstract. The Self-Organising Map (SOM) is a well-known neuralnetwork model that has successfully been used as a data analysis tool in many different domains. The SOM provides a topology-preserving mapping from a high-dimensional input space to a lower-dimensional output space, a convenient interface to the data. However, the real power of this model can only be utilised with sophisticated visualisations that provide a powerful tool-set for exploring and understanding the characteristics of the underlying data. We thus present a novel visualisation technique that is able to illustrate the structure inherent in the data. The method builds on minimum spanning trees as a graph of similar data items, which is subsequently visualised on top of the SOM grid.

show abstract

Advanced visualization of Self-Organizing Maps with vector fields

Cited by 38 publications

References 22 publications

Analysing a large dataset on long-term monitoring of water quality and plankton with the SOM clustering

Analysing a large dataset on long-term monitoring of water quality and plankton with the SOM clustering

Multiscale visual quality assessment for cluster analysis with self-organizing maps

Visualising Clusters in Self-Organising Maps with Minimum Spanning Trees

Contact Info

Product

Resources

About