Self-organizing maps are a prominent unsupervised neural network model providing cluster analysis of highdimensional input data. However, in spite of enhanced visualization techniques for self-organizing maps, interpreting a trained map proves to be difficult because the features responsible for a specific cluster assignment are not evident from the resulting map representation. In this paper we present our LabelSOM approach for automatically labeling a trained selforganizing map with the features of the input data that are the most relevant ones for the assignment of a set of input data to a particular cluster. The resulting labeled map allows the user to understand the structure and the information available in the map and the reason for a specific map organization, especially when only little prior information on the data set and its characteristics is available. We demonstrate the applicability of the LabelSOM method in the field of data mining providing an example from real world...

. Digital Libraries have gained tremendous interest with several research projects addressing the wealth of challenges in this field. While computational intelligence systems are being used for specific tasks in this arena, the majority of projects relies on conventional techniques for the basic structure of the library itself. With the SOMLib project we created a digital library system that uses a neural network-based core for the representation of the library. The self-organizing map, a popular unsupervised neural network model, is used to topically structure a document collection similar to the organization of real-world libraries. Based on this core, additional modules provide information retrieval features, integrate distributed libraries, and automatically label the various topical sections in the document collection. A metaphor graphics based interface further assists the user in intuitively understanding the library providing an instant overview. Keywords: Self-Organizing Map ...

. Self-organizing maps are an unsupervised neural network model which lends itself to the cluster analysis of high-dimensional input data. However, interpreting a trained map proves to be difficult because the features responsible for a specific cluster assignment are not evident from the resulting map representation. In this paper we present our LabelSOM approach for automatically labeling a trained self-organizing map with the features of the input data that are the most relevant ones for the assignment of a set of input data to a particular cluster. The resulting labeled map allows the user to better understand the structure and the information available in the map and the reason for a specific map organization, especially when only little prior information on the data set and its characteristics is available. 1 Introduction The self-organizing map (SOM) [2, 3] is a prominent unsupervised neural network model for cluster analysis. Data from a high-dimensional input space is mapped...

The self-organizing map is a very popular unsupervised neural network model for the analysis of high-dimensional input data as in information retrieval applications. However, the interpretation of the map requires much manual eort, especially as far as the analysis of the learned features and the characteristics of identi ed clusters is concerned. In this paper we present the LabelSOM method which, based on the features learned by the map, automatically selects the most descriptive features of the input patterns mapped onto a particular unit of the map, thus making the characteristics of the various clusters within the map explicit. We demonstrate the bene ts of this approach on an example from text classi cation using a real-world document archive. In this particular case, the features correspond to keywords describing the contents of a document. The bene t of this approach is that the various document clusters are characterized in terms of shared keywords, thus making it easy for the user to explore the contents of an unknown document archive.

While methods for searching large digital libraries have experienced tremendous improvements recently, interfaces to such collections still have a far way to go. Most interfaces to digital libraries present themselves as various forms of sorted lists, providing metadata information on the documents in textual form. This prohibits intuitive understanding of document archives or web search results. In this paper we present the libViewer, a Java-based user interface to digital libraries using metaphor graphics to display information on the elements in a digital library in an intuitively understandable way. Metadata on digital libraries based on the Dublin Core Initiative is mapped onto a set of metaphors to allow instant recognition and orientation in an unknown document collection, facilitating interactive retrieval and exploration. Keywords: Information Visualization, Web-based Library Representation, Information Space Metaphors 1. Introduction When entering a library or large-scale b...

While methods for retrieving documents from large information repositories have improved a lot, presentation of the retrieved documents still leaves a lot to be desired. Important information on documents is usually presented as a textual listing of available metadata attributes such as document size, author information, date of creation, and so on. This requires the user to read and abstract from the presented metainformation. In this paper we present our libViewer system, a Java-Applet interfacing with a number of servers to provide an intuitive metaphor-graphics based representation of document repositories. Contrary to most other multidimensional data visualization approaches we rely on intuitive realworld metaphors to provide a visualization for untrained users rather than experts in special interfaces. We introduce a set of metaphors and present two prototype systems interfacing with Dublin Core metadata based repositories as well as the AltaVista search engine. We further provid...

. While the focus of research concerning electronic document archives still is on information retrieval, the importance of interactive exploration has been realized and is gaining importance. The map metaphor, where documents are organized on a map according to their contents, has proven particularly useful as an interface to such a collection. The self-organizing map has shown to produce stable topically ordered organizations of documents on such a 2-dimensional map display. However, the characteristics of these topical clusters are not being made explicit. In this paper we present the LabelSOM method which takes the applicability of the self-organizing map for document archive organization one step further by automatically labeling the various topical clusters found in the map. This allows the user to get an instant overview of the various topics covered by a document collection. 1 Introduction Today's information age may be characterized by constant massive production and dissemina...

The self-organizing map is a very popular unsupervised neural network model for the analysis of high-dimensional input data as it is typically found in information retrieval applications. However, the interpretation of the map requires much manual effort, especially as far as the analysis of the learned features and the characteristics of identified clusters is concerned. In this paper we present our novel LabelSOM method which, based on the features learned by the map, automatically selects the most descriptive features of the input patterns mapped onto a particular unit of the map, thus making the associations between the various clusters within the map explicit. We demonstrate the benefits of this approach with examples from text classification using two different real-world document archives. In this particular case, the features correspond to keywords describing the contents of a document. The benefit of this approach is obvious in that the various document clusters are character...

Text collections may be regarded as an almost perfect application arena for unsupervised neural networks. This is because many operations computers have to perform on text documents are classification tasks based on noisy patterns. In particular we rely on self-organizing maps which produce a map of the document space after their training process. From geography, however, it is known that maps are not always the best way to represent information spaces. For most applications it is better to provide a hierarchical view of the underlying data collection in form of an atlas where, starting from a map representing the complete data collection, different regions are shown at finer levels of granularity. Using an atlas, the user can easily "zoom" into regions of particular interest while still having general maps for overall orientation. We show that a similar display can be obtained by using hierarchical feature maps to represent the contents of a document archive. These neural networks have layerd architecture where each layer consists of a number of individual self-organizing maps. By this, the contents of the text archive may be represented at arbitrary detail while still having the general maps available for global orientation.

With the increasing amount of information available in electronic document collections, methods for organizing these collections to allow topic-oriented browsing and orientation gain increasing importance. In this paper, we present the Growing Hierarchical Self-Organizing Map, which allows an automatic hierarchical decomposition and organization of documents. We present a case study based on a 3-month article collection from an Austrian daily newspaper.