Search results

The aim of this paper is to explore the possibility of retrieving information with Kohonen self‐organising maps, which are known to be effective to group objects according to their similarity or dissimilarity.

After conventional preprocessing, such as transforming into vector space, documents from a German document collection were trained for a neural network of Kohonen self‐organising map type. Such an unsupervised network forms a document map from which relevant objects can be found according to queries.

Self‐organising maps ordered documents to groups from which it was possible to find relevant targets.

The number of documents used was moderate due to the limited number of documents associated to test topics. The training of self‐organising maps entails rather long running times, which is their practical limitation. In future, the aim will be to build larger networks by compressing document matrices, and to develop document searching in them.

With self‐organising maps the distribution of documents can be visualised and relevant documents found in document collections of limited size.

The paper reports on an approach that can be especially used to group documents and also for information search. So far self‐organising maps have rarely been studied for information retrieval. Instead, they have been applied to document grouping tasks.

Studies aspects of Heinz von Foerster's work that are of particular importance for cognitive science and artificial intelligence.

Kohonen's self‐organizing map is presented as one method that may be useful in implementing some of Von Foerster's ideas. The main foci are the distinction between trivial and non‐trivial machines and the concept of constructive learning. The self‐organizing map is also presented as a potential tool for alleviating the participatory crisis discussed by von Foerster.

The participatory crisis in society is discussed and the concept of change is handled within the framework of information systems development.

Considers the importance of considering change in information systems development.

The goal of the research is to explore whether the use of higher‐level semantic features can help us to build better self‐organising map (SOM) representation as measured from a human‐centred perspective. The authors also explore an automatic evaluation method that utilises human expert knowledge encapsulated in the structure of traditional textbooks to determine map representation quality.

Two types of document representations involving semantic features have been explored – i.e. using only one individual semantic feature, and mixing a semantic feature with keywords. Experiments were conducted to investigate the impact of semantic representation quality on the map. The experiments were performed on data collections from a single book corpus and a multiple book corpus.

Combining keywords with certain semantic features achieves significant improvement of representation quality over the keywords‐only approach in a relatively homogeneous single book corpus. Changing the ratios in combining different features also affects the performance. While semantic mixtures can work well in a single book corpus, they lose their advantages over keywords in the multiple book corpus. This raises a concern about whether the semantic representations in the multiple book corpus are homogeneous and coherent enough for applying semantic features. The terminology issue among textbooks affects the ability of the SOM to generate a high quality map for heterogeneous collections.

The authors explored the use of higher‐level document representation features for the development of better quality SOM. In addition the authors have piloted a specific method for evaluating the SOM quality based on the organisation of information content in the map.

This paper aims to analyze the relationship among measures of resource and service usage and other features of academic libraries in the USA and Canada.

Through the use of a self‐organizing map, academic library data were clustered and visualized. Analysis of the library data was conducted through the computation of a “library performance metric” that was applied to the resulting map.

Two areas of high‐performing academic libraries emerged on the map. One area included libraries with large numbers of resources, while another area included libraries that had low resources but gave greater numbers of presentations to groups, offered greater numbers of public service hours, and had greater numbers of staffed service points.

The metrics chosen as a measure of library performance offer only a partial picture of how libraries are being used. Future research might involve the use of a self‐organizing map to cluster library data within certain parameters and the identification of high‐performing libraries within these clusters.

This study suggests that libraries can improve their performance not only by acquiring greater resources but also by putting greater emphasis on the services that they provide to their users.

This paper demonstrates how a self‐organizing map can be used in the analysis of large data sets to facilitate library comparisons.

Multilevel environment analysis is important for companies operating on the global market. Previous studies have in general focused on one level at a time, but the need to perform multilevel environment analysis has also been stressed. Multilevel analysis can partly explain the benchmarking gap between companies, as changing conditions in the upper environment levels affect lower levels. In today's information‐rich era, it is difficult to conduct multilevel analysis without suitable computational tools. This paper illustrates how the self‐organizing map can be used for the simultaneous comparison of industry‐level changes and financial performance of pulp and paper companies. The study shows the importance of simultaneous analysis, as some simultaneous changes were found at both industry and corporate levels. Also found were some industry‐specific explanatory factors for good (Scandinavian companies) and poor (Japanese companies) financial performance. The results indicate that the self‐organizing map could be a suitable tool when the purpose is to visualize large masses of multilevel data from high‐dimensional databases.

The purpose of this paper is to examine neural document clustering techniques, e.g. self‐organising map (SOM) or growing neural gas (GNG), usually assume that textual information is stationary on the quantity.

The authors propose a novel dynamic adaptive self‐organising hybrid (DASH) model, which adapts to time‐event news collections not only to the neural topological structure but also to its main parameters in a non‐stationary environment. Based on features of a time‐event news collection in a non‐stationary environment, they review the main current neural clustering models. The main deficiency is a need of pre‐definition of the thresholds of unit‐growing and unit‐pruning. Thus, the dynamic adaptive self‐organising hybrid (DASH) model is designed for a non‐stationary environment.

The paper compares DASH with SOM and GNG based on an artificial jumping corner data set and a real world Reuters news collection. According to the experimental results, the DASH model is more effective than SOM and GNG for time‐event document clustering.

A real world environment is dynamic. This paper provides an approach to present news clustering in a non‐stationary environment.

Text clustering in a non‐stationary environment is a novel concept. The paper demonstrates DASH, which can deal with a real world data set in a non‐stationary environment.

Image analysis faces data reduction problems when deriving low‐dimensional image spaces (‘perceptual maps’) from multidimensional profile data. The neurocomputing methodology of Self‐Organizing Maps may contribute to finding a radically parsimonious representation. The principles of SOM methodology are shown in a case study on the company images of nine Austrian tour operators.

Browsing knowledge documents in a peer-to-peer (P2P) environment is difficult because knowledge documents in such an environment are large in quantity and distributed over different peers who organize the documents according to their own views. This paper aims to propose a method for constructing a personal knowledge map for a peer to facilitate knowledge browsing and alleviate information overload in P2P environments.

The research presents a method for constructing a personal knowledge map. The method adopts an ontology-concept-tree-based classification algorithm to recognize a peer’s personal knowledge structure and construct a personal knowledge map, and uses a self-organizing map algorithm to cluster and visualize the knowledge documents. The correctness of the created knowledge map is evaluated with a collection of abstracts of academic papers.

The method for constructing a personal knowledge map is the main finding of this research. The evaluation shows that the created knowledge map is good in quality.

The proposed method provides a way for P2P platforms to understand their users’ knowledge background, as well as to improve the P2P platform environment. However, the proposed method will not help a peer when he has nothing in his individual knowledge document repository (i.e. the “cold start” problem). The method also requires a relatively good ontology base for a P2P document sharing system to use the method effectively.

It is novel that the proposed method organizes the knowledge documents related to a peer’s knowledge background into a personal knowledge map. Moreover, the created knowledge map combines the advantages of a hierarchical display and a map display. It has values for a distributed P2P environment to facilitate users’ knowledge browsing and to alleviate information overload.

The Tourism Knowledge Map is part of a larger project that develops a web portal entitled The Tourism Knowledge Base. On this web site the users will be offered comprehensive information about the organizations in Austria providing tourism education, research, and consulting services. The Knowledge Map assists the users in finding and optimizing a set of keywords for launching an efficient search operation in tourism‐centred databases accessible on the internet. The underlying method is the Self‐Organizing Map, one of the most widely accepted techniques of unsupervised learning. Three real‐world examples illustrate how a Knowledge Map may be constructed from the frequencies of keywords and their co‐occurrence in the abstracts of tourism‐related research papers, articles, and reports.

According to the problems of visual perception, we propose a model for the processing of vision in adverse situations of illumination, scale, etc. In this paper, a model for image segmentation and labelling obtained in real conditions with different scales is proposed.

The model is based on the texture identification of the scene's objects by means of comparison with a database that stores series of each texture perceived with successive optic parameter values. As a basis for the model, self‐organising maps have been used in several phases of the labelling process.

The model has been conceived to systematically deal with the different causes that make vision difficult and allows it to be applied in a wide range of real situations. The results show high success rates in the labelling of scenes captured in different scale conditions, using very simple describers, such as different histograms of textures.

Our interest is directed towards systematising the proposal and experimenting on the influence of the other variables of the vision. We will also tackle the implantation of the classifier module so that the different causes can be dealt with by the reconfiguration of the same hardware (using reconfigurable hardware).

This research approaches a very advanced angle of the vision problems: visual perception under adverse conditions. In order to deal with this problem, a model formulated with a general purpose is proposed. Our objective is to present an approach to conceive universal architectures (in the sense of being valid with independence of the implied magnitudes).