This article outlines similarity applied to the general environment and geographical information domains. The hypothesis is if physical and social sciences manifest similar amenities, then similarity would be a generative technique to analyse the cached information inherent in the data retrieved. Similarity is examined concerning the spatial grouping of natural kinds in a complex environment.
Download (gzipped PDF, 165KB)
An approach is presented for incorporating metatata constraints into queries to be processed by a distributed environmental information system. The approach, based on a novel metamodel unifying concepts from the Unified Modelling Language (UML), the Object Query Language (OQL), and the Resource Description Framework (RDF), allows metadata information to be represented and processed in combination with regular data queries.
Download (gzipped PDF, 126KB)
We report on the clustering of nodes in internally represented acoustic space. Learners of different languages partition perceptual space distinctly. Here, an Evolving Connectionist-Based System (ECOS) is used to model the perceptual space of New Zealand English. Currently, the system evolves in an unsupervised, self-organising manner. The perceptual space can be visualised, and the important features of the input patterns analysed. Additionally, the path of the internal representations can be seen. The results here will be used to develop a supervised system that can be used for speech recognition based on the evolved, internal sub-word units.
Download (gzipped PDF, 117KB)
The English and Mäori word translator ngä aho whakamäori-ä-tuhi was designed to provide single head-word translations to on-line web users. There are over 13,000 words all based on traditional text sources, derived because of their high frequency used within each of the respective languages. The translator has been operational for well over a year now, and it has had the highest web traffic usage in the Department of Information Science. Two log files were generated to record domain hits and language translations, both provided the up-to-date data for analysis contained in this paper.
Download (gzipped PDF, 183KB)
Mäori speech data collection and analysis is an ongoing process, as new and existing data sets are continuously accessed for many different experimental speech perception and generation processing tasks. A data management system is an important tool to facilitate the systematic techniques applied to the speech and language data. Identification of the core components for Mäori speech and language databases, translation systems, speech recognition and speech synthesis have been undertaken as research themes. The latter component will be the main area of discussion here. So to hasten the development of Mäori speech synthesis, joint collaborative research with established international projects has begun. This will allow the Mäori language to be presented to the wider scientific community well in advance of other similar languages, many times it’s own size and distribution. Propagation of the Mäori language via the information communication technology (ICT) medium is advantageous to it’s long term survival.
Download (gzipped PDF, 228KB)
In real world information systems, data analysis and processing are usually needed to be done in an on-line, self-adaptive way. In this respect, neural algorithms of incremental learning and constructive network models are of increased interest. In this paper we present a new algorithm of evolving self-organizing map (ESOM), which features fast one-pass learning, dynamic network structure, and good visualisation ability. Simulations have been carried out on some benchmark data sets for classification and prediction tasks, as well as on some macroeconomic data for data analysis. Compared with other methods, ESOM achieved better classification with much shorter learning time. Its performance for time series modelling is also comparable, requiring more hidden units but with only one-pass learning. Our results demonstrate that ESOM is an effective computational model for on-line learning, data analysis and modelling.
Download (gzipped PDF, 178KB)
It is canonical practice in agent-based systems to use a declarative format for the exchange of information. The increasing usage and facility of object-oriented tools and techniques, however, suggests there may be benefits in combining the use of object-oriented modelling approaches with agent-based messaging. In this paper we outline our efforts in connection with the New Zealand Distributed Information Systems project to use object-oriented knowledge representation in an agent-based architecture. Issues and tradeoffs are discussed, as well as the possible extensions to current agent-based message protocols that may be necessary in order to support object-oriented information exchange.
Download (gzipped PDF, 69KB)
This paper describes a system of interlinked ontologies to describe the concepts underlying FIPA agent communication. A meta-modelling approach is used to relate object-oriented domain ontologies and abstract models of agent communication and content languages and to describe them in a single framework. The modelling language used is the Unified Modeling Language, which is extended by adding the concepts of resource and reference. The resulting framework provides an elegant basis for the development of agent systems that combine object-oriented information representation with agent messaging protocols.
Download (gzipped PDF, 78KB)
When modelling a real-world phenomenon, it can often be useful to have multiple descriptions of the phenomenon, each expressed using a different modelling approach or representation. Different representations such as entity-relationship modelling, data flow modelling and use case modelling allow analysts to describe different aspects of real-world phenomena, thus providing a more thorough understanding than if a single representation were used. Researchers working with multiple representations have approached the problem from many different fields, resulting in a diverse and potentially confusing set of terminologies. In this paper is described a viewpoint-based framework for discussing the use of multiple modelling representations to describe real-world phenomena. This framework provides a consistent and integrated terminology for researchers working with multiple representations. An abstract notation is also defined for expressing concepts within the framework.
Download (gzipped PDF, 377KB)
Most applications of Genetic Programming to time series modeling use a fitness measure for comparing potential solutions that treat each point in the time series independently. This non-temporal approach can lead to some potential solutions being given a relatively high fitness measure even though they do not correspond to the training data when the overall shape of the series is taken into account. This paper develops two fitness measures which emphasize the concept of shape when measuring the similarity between a training and evolved time series. One approach extends the root mean square error to higher dimensional derivatives of the series. The second approach uses a simplified derivative concept that describes shape in terms of positive, negative and zero slope.
Download (gzipped PDF, 432KB)
An important part of the systems development process is building models of real-world phenomena. These phenomena are described by many different kinds of information, and this diversity has resulted in a wide variety of modelling representations. Some types of information are better expressed by some representations than others, so it is sensible to use multiple representations to describe a real-world phenomenon. The author has developed an approach to facilitating the use of multiple representations within a single viewpoint by translating descriptions of the viewpoint among different representations. An important issue with such translations is their quality, or how well they map constructs of one representation to constructs of another representation. Two possible methods for improving translation quality, heuristics and enrichment, are proposed in this paper, and a preliminary metric for measuring relative translation quality is described.
Download (gzipped PDF, 376KB)
Workflow management systems are increasingly used to assist the automation of business processes that involve the exchange of documents, information, or task execution results. Recent developments in distributed information system technology now make it possible to extend the workflow management system idea to much wider spheres of activity in the industrial and commercial world. This paper describes a framework under development that employs such technology so that software tools and processes may interoperate in a distributed and dynamic environment. Key technical elements of the framework include the use of coloured Petri nets and distributed object technology (CORBA).
Keywords: distributed systems, workflow, process modelling
Download (gzipped PDF, 134KB)
The use of modelling abstractions to map from items in the real-world to objects in the computational domain is useful both for the effective implementation of abstract problem solutions and for the management of software complexity. This paper discusses the new approach of agent-oriented software engineering (AOSE), which uses the notion of an autonomous agent as its fundamental modelling abstraction. For the AOSE approach to be fully exploited, software engineers must be able to gain leverage from an agent software architecture and framework, and there are several such frameworks now publicly available. At the present time, however, there is little information concerning the options that are available and what needs to be considered when choosing or developing an agent framework. We consider three different agent software architectures that are (or will be) publicly available and evaluate some of the design and architectural differences and trade-offs that are associated with them and their impact on agent-oriented software development. Our discussion examines these frameworks in the context of an example in the area of distributed information systems.
Download (gzipped PDF, 207KB)
Electronic medical consultation is available worldwide through access to the World Wide Web (WWW). This article outlines a research study on the adoption of electronic medical consultation as a means of health delivery. It focuses on the delivery of healthcare specifically for New Zealanders, by New Zealanders. It is acknowledged that the WWW is a global market place and it is therefore difficult to identify New Zealanders’ use of such a global market, but we have attempted to provide a New Zealand perspective on electronic medical consultation.
Keywords: electronic medical consultation, on-line health, New Zealand
Download (gzipped PDF, 60KB)
The paper applies novel techniques for on-line, adaptive learning of macroeconomic data and a consecutive analysis and prediction. The evolving connectionist system paradigm (ECOS) is used in its two versions—unsupervised (evolving self-organised maps), and supervised (evolving fuzzy neural networks—EFuNN). In addition to these techniques self-organised maps (SOM) are also employed for finding clusters of countries based on their macroeconomic parameters. EFuNNs allow for modelling, clustering, prediction and rule extraction. The rules that describe future annual values for the consumer price index (CPI), interest rate, unemployment and GDP per capita are extracted from data and reported in the paper for both global—EU-Asia block of countries, and for smaller groups—EU, EU-candidate countries, Asia-Pacific countries. The analysis and prediction models proof to be useful tools for the analysis of trends in macroeconomic development of clusters of countries and their future prediction.
Keywords: neural networks, fuzzy rules, evolving connectionist systems, macroeconomic clusters
Although widely studied for many years, colour image quantisation remains a challenging problem. We propose to use an evolving self-organising map model for the on-line image quantisation tasks. Encouraging results are obtained in experiments and we look forward to implementing the algorithm in real world applications with further improvement.
Download (gzipped PDF, 2.1MB)
In a consensus ensemble all members must agree before they classify a data point. But even when they all agree some data is still misclassified. In this paper we look closely at consistently misclassified data to investigate whether some of it may be outliers or may have been mislabeled.
Download (gzipped PDF, 198KB)
Elsewhere Kennedy describes three elementary structures to be found in entity-relationship diagrams. Here, each of these structures is considered in the context of a transaction processing system and a specific set of components that can be associated with the structure is described. Next, an example is given illustrating the use of elementary structures as an analytical tool for data modelling and a diagnostic tool for the identification of errors in the resulting data model. It is conjectured that the amount of effort associated with each structure can be measured. A new approach for the estimation of the total effort required to develop a system, based on a count of the elementary structures present in the entity-relationship diagram, is then proposed. The approach is appealing because it can be automated and because it can be applied earlier in the development cycle than other estimation methods currently in use. The question of a suitable counting strategy remains open.
Keywords: data modelling, design tools and techniques, entity-relationship model, software metrics
Download (gzipped PDF, 81KB)
In any data set there some of the data will be bad or noisy. This study identifies two types of noise and investigates the effect of each in the training data of backpropagation neural networks. It also compares the mean square error function with a more robust alternative advocated by Huber.
Download (gzipped PDF, 198KB)
Workflow management systems (WFMS) are being adopted to assist the automation of business processes that involve the exchange of information. As a result of developments in distributed information system technology, it is now possible to extend the WFMS idea to wider spheres of activity in the industrial and commercial world and thereby to encompass the increasingly sprawling nature of modern organisations. This paper describes a framework under development that employs such technology so that software tools and processes may interoperate in a distributed and dynamic environment. The framework employs Petri nets to model the interaction between various sub-processes. CORBA technology is used to enable different participants who are physically disparate to monitor activity in and make resource-level adaptations to their particular subnet.
Keywords: distributed systems, workflow, process modelling, Petri nets
Download (gzipped PDF, 199KB)