63 results for Working or discussion paper, 1999

  • A practical guide to GMM (with applications to option pricing)

    Arnold, Tom; Crack, Timothy (1999-10)

    Working or discussion paper
    University of Otago

    The full text of this document is only available from the Social Science Research Network. Please use the related link to access the full text.

    View record details
  • Dynamic evolving fuzzy neural networks with `m-out-of-n' activation nodes for on-line adaptive systems

    Kasabov, Nikola; Song, Qun (1999-03)

    Working or discussion paper
    University of Otago

    The paper introduces a new type of evolving fuzzy neural networks (EFuNNs), denoted as mEFuNNs, for on-line learning and their applications for dynamic time series analysis and prediction. mEFuNNs evolve through incremental, hybrid (supervised/unsupervised), on-line learning, like the EFuNNs. They can accommodate new input data, including new features, new classes, etc. through local element tuning. New connections and new neurons are created during the operation of the system. At each time moment the output vector of a mEFuNN is calculated based on the m-most activated rule nodes. Two approaches are proposed: (1) using weighted fuzzy rules of Zadeh-Mamdani type; (2) using Takagi-Sugeno fuzzy rules that utilise dynamically changing and adapting values for the inference parameters. It is proved that the mEFuNNs can effectively learn complex temporal sequences in an adaptive way and outperform EFuNNs, ANFIS and other neural network and hybrid models. Rules can be inserted, extracted and adjusted continuously during the operation of the system. The characteristics of the mEFuNNs are illustrated on two bench-mark dynamic time series data, as well as on two real case studies for on-line adaptive control and decision making. Aggregation of rule nodes in evolved mEFuNNs can be achieved through fuzzy C-means clustering algorithm which is also illustrated on the bench mark data sets. The regularly trained and aggregated in an on-line, self-organised mode mEFuNNs perform as well, or better, than the mEFuNNs that use fuzzy C-means clustering algorithm for off-line rule node generation on the same data set.

    View record details
  • Software metrics data analysis—Exploring the relative performance of some commonly used modeling techniques

    Gray, Andrew; MacDonell, Stephen (1999-06)

    Working or discussion paper
    University of Otago

    Whilst some software measurement research has been unquestionably successful, other research has struggled to enable expected advances in project and process management. Contributing to this lack of advancement has been the incidence of inappropriate or non-optimal application of various model-building procedures. This obviously raises questions over the validity and reliability of any results obtained as well as the conclusions that may have been drawn regarding the appropriateness of the techniques in question. In this paper we investigate the influence of various data set characteristics and the purpose of analysis on the effectiveness of four model-building techniques---three statistical methods and one neural network method. In order to illustrate the impact of data set characteristics, three separate data sets, drawn from the literature, are used in this analysis. In terms of predictive accuracy, it is shown that no one modeling method is best in every case. Some consideration of the characteristics of data sets should therefore occur before analysis begins, so that the most appropriate modeling method is then used. Moreover, issues other than predictive accuracy may have a significant influence on the selection of model-building methods. These issues are also addressed here and a series of guidelines for selecting among and implementing these and other modeling techniques is discussed.

    View record details
  • Spatial-temporal adaptation in evolving fuzzy neural networks for on-line adaptive phoneme recognition

    Kasabov, Nikola; Watts, Michael (1999-03)

    Working or discussion paper
    University of Otago

    Please note that this is a searchable PDF derived via optical character recognition (OCR) from the original source document. As the OCR process is never 100% perfect, there may be some discrepancies between the document image and the underlying text. Searching and selecting the text of this PDF may also not work in all viewers; for example, they have been found to not work in Apple's Preview application. We therefore recommend Adobe Reader for viewing and searching this PDF.

    View record details
  • UML as an ontology modelling language

    Cranefield, Stephen; Purvis, Martin (1999-01)

    Working or discussion paper
    University of Otago

    Current tools and techniques for ontology development are based on the traditions of AI knowledge representation research. This research has led to popular formalisms such as KIF and KL-ONE style languages. However, these representations are little known outside AI research laboratories. In contrast, commercial interest has resulted in ideas from the object-oriented programming community maturing into industry standards and powerful tools for object-oriented analysis, design and implementation. These standards and tools have a wide and rapidly growing user community. This paper examines the potential for object-oriented standards to be used for ontology modelling, and in particular presents an ontology representation language based on a subset of the Unified Modeling Language together with its associated Object Constraint Language.

    View record details
  • From hybrid adjustable neuro-fuzzy systems to adaptive connectionist-based systems for phoneme and word recognition

    Kasabov, Nikola; Kilgour, Richard; Sinclair, Stephen (1999-03)

    Working or discussion paper
    University of Otago

    This paper discusses the problem of adaptation in automatic speech recognition systems (ASRS) and suggests several strategies for adaptation in a modular architecture for speech recognition. The architecture allows for adaptation at different levels of the recognition process, where modules can be adapted individually based on their performance and the performance of the whole system. Two realisations of this architecture are presented along with experimental results from small-scale experiments. The first realisation is a hybrid system for speaker-independent phoneme-based spoken word recognition, consisting of neural networks for recognising English phonemes and fuzzy systems for modelling acoustic and linguistic knowledge. This system is adjustable by additional training of individual neural network modules and tuning the fuzzy systems. The increased accuracy of the recognition through appropriate adjustment is also discussed. The second realisation of the architecture is a connectionist system that uses fuzzy neural networks FuNNs to accommodate both a prior linguistic knowledge and data from a speech corpus. A method for on-line adaptation of FuNNs is also presented.

    View record details
  • Industry practices in project management for multimedia information systems

    MacDonell, Stephen; Fletcher, Tim (1999-06)

    Working or discussion paper
    University of Otago

    This paper describes ongoing research directed at formulating a set of appropriate measures for assessing and ultimately predicting effort requirements for multimedia systems development. Whilst significant advances have been made in the determination of measures for both transaction-based and process-intensive systems, very little work has been undertaken in relation to measures for multimedia systems. A small preliminary empirical study is reviewed as a precursor to a more exploratory investigation of the factors that are considered by industry to be influential in determining development effort. This work incorporates the development and use of a goal-based framework to assist the measure selection process from a literature basis, followed by an industry questionnaire. The results provide a number of preliminary but nevertheless useful insights into contemporary project management practices with respect to multimedia systems.

    View record details
  • A distributed architecture for environmental information systems

    Purvis, Martin; Cranefield, Stephen; Nowostawski, Mariusz (1999-04)

    Working or discussion paper
    University of Otago

    The increasing availability and variety of large environmental data sets is opening new opportunities for data mining and useful cross-referencing of disparate environmental data sets distributed over a network. In order to take advantage of these opportunities, environmental information systems will need to operate effectively in a distributed, open environment. In this paper, we describe the New Zealand Distributed Information System (NZDIS) software architecture for environmental information systems. In order to optimise extensibility, openness, and flexible query processing, the architecture is organised into collaborating software agents that communicate by means of a standard declarative agent communication language. The metadata of environmental data sources are stored as part of agent ontologies, which represent information models of the domain of the data repository. The agents and the associated ontological framework are designed as much as possible to take advantage of standard object-oriented technology, such as CORBA, UML, and OQL, in order to enhance the openness and accessibility of the system.

    View record details
  • Finding medical information on the Internet: Who should do it and what should they know

    Parry, David (1999-06)

    Working or discussion paper
    University of Otago

    More and more medical information is appearing on the Internet, but it is not easy to get at the nuggets amongst all the spoil. Bruce McKenzie's editorial in the December 1997 edition of SIM Quarterly dealt very well with the problems of quality, but I would suggest that the problem of accessibility is as much of a challenge. As ever-greater quantities of high quality medical information are published electronically, the need to be able to find it becomes imperative. There are a number of tools to find what you want on the Internet---search engines, agents, indexing and classification schemes and hyperlinks, but their use requires care, skill and experience.

    View record details
  • Adaptive, evolving, hybrid connectionist systems for image pattern recognition

    Kasabov, Nikola; Israel, Steven; Woodford, Brendon J (1999-05)

    Working or discussion paper
    University of Otago

    Please note that this is a searchable PDF derived via optical character recognition (OCR) from the original source document. As the OCR process is never 100% perfect, there may be some discrepancies between the document image and the underlying text. Searching and selecting the text of this PDF may also not work in all viewers; for example, they have been found to not work in Apple's Preview application. We therefore recommend Adobe Reader for viewing and searching this PDF.

    View record details
  • Infiltrating IT into primary care: A case study

    Cockcroft, Sophie; Parry, David; Breton, Alice; Abernethy, David; Gillies, John (1999-12)

    Working or discussion paper
    University of Otago

    Web based approaches to tracking students on placement are receiving much interest in the field of medical education The work presented here describes a web-based solution to the problem of managing data collection from student encounters with patients whilst on placement. The solution has been developed by postgraduate students under the direction of staff of the health informatics diploma. Specifically, the system allows undergraduate students on placement or in the main hospital to access a web-based front end to a database designed to store the data that they are required to gather. The system also has the important effect of providing a rationale for the provision of electronic communication to the undergraduate students within the context of healthcare delivery. We believe that an additional effect will be to expose practicing healthcare providers to electronic information systems, along with the undergraduates who are trained to use them, and increase the skill base of the practitioners.

    View record details
  • The concepts of hidden Markov model in speech recognition

    Abdulla, Waleed H; Kasabov, Nikola (1999-05)

    Working or discussion paper
    University of Otago

    The speech recognition field is one of the most challenging fields that has faced scientists for a long time. The complete solution is still far from reach. The efforts are concentrated with huge funds from the companies to different related and supportive approaches to reach the final goal. Then, apply it to the enormous applications that are still waiting for the successful speech recognisers that are free from the constraints of speakers, vocabularies or environment. This task is not an easy one due to the interdisciplinary nature of the problem and as it requires speech perception to be implied in the recogniser (Speech Understanding Systems) which in turn point strongly to the use of intelligence within the systems. The bare techniques of recognisers (without intelligence) are following wide varieties of approaches with different claims of success by each group of authors who put their faith in their favourite way. However, the sole technique that gains the acceptance of the researchers to be the state of the art is the Hidden Markov Model (HMM) technique. HMM is agreed to be the most promising one. It might be used successfully with other techniques to improve the performance, such as hybridising the HMM with Artificial Neural Networks (ANN) algorithms. This does not mean that the HMM is pure from approximations that are far from reality, such as the successive observations independence, but the results and potential of this algorithm is reliable. The modifications on HMM take the burden of releasing it from these poorly representative approximations hoping for better results. In this report we are going to describe the backbone of the HMM technique with the main outlines for successful implementation. The representation and implementation of HMM varies in one way or another but the main idea is the same as well as the results and computation costs, it is a matter of preferences to choose one. Our preference here is that adopted by Ferguson and Rabiner et al. In this report we will describe the Markov Chain, and then investigate a very popular model in the speech recognition field (the Left-Right HMM Topology). The mathematical formulations needed to be implemented will be fully explained as they are crucial in building the HMM. The prominent factors in the design will also be discussed. Finally we conclude this report by some experimental results to see the practical outcomes of the implemented model.

    View record details
  • Factors systematically associated with errors in subjective estimates of software development effort: The stability of expert judgment

    Gray, Andrew; MacDonell, Stephen; Shepperd, Martin (1999-06)

    Working or discussion paper
    University of Otago

    Software metric-based estimation of project development effort is most often performed by expert judgment rather than by using an empirically derived model (although such may be used by the expert to assist their decision). One question that can be asked about these estimates is how stable are they with respect to characteristics of the development process and product? This stability can be assessed in relation to the degree to which the project has advanced over time, the type of module for which the estimate is being made, and the characteristics of that module. In this paper we examine a set of expert-derived estimates for the effort required to develop a collection of modules from a large health-care system. Statistical tests are used to identify relationships between the type (screen or report) and characteristics of modules and the likelihood of the associated development effort being under-estimated, approximately correct, or over-estimated. Distinct relationships are found that suggest that the estimation process being examined was not unbiased to such characteristics.

    View record details
  • Assessing prediction systems

    Kitchenham, Barbara; MacDonell, Stephen; Pickard, Lesley; Shepperd, Martin (1999-06)

    Working or discussion paper
    University of Otago

    For some years software engineers have been attempting to develop useful prediction systems to estimate such attributes as the effort to develop a piece of software and the likely number of defects. Typically, prediction systems are proposed and then subjected to empirical evaluation. Claims are then made with regard to the quality of the prediction systems. A wide variety of prediction quality indicators have been suggested in the literature. Unfortunately, we believe that a somewhat confusing state of affairs prevails and that this impedes research progress. This paper aims to provide the research community with a better understanding of the meaning of, and relationship between, these indicators. We critically review twelve different approaches by considering them as descriptors of the residual variable. We demonstrate that the two most popular indicators MMRE and pred(25) are in fact indicators of the spread and shape respectively of prediction accuracy where prediction accuracy is the ratio of estimate to actual (or actual to estimate). Next we highlight the impact of the choice of indicator by comparing three prediction systems derived using four different simulated datasets. We demonstrate that the results of such a comparison depend upon the choice of indicator, the analysis technique, and the nature of the dataset used to derive the predictive model. We conclude that prediction systems cannot be characterised by a single summary statistic. We suggest that we need indicators of the central tendency and spread of accuracy as well as indicators of shape and bias. For this reason, boxplots of relative error or residuals are useful alternatives to simple summary metrics.

    View record details
  • Using the Internet to teach health informatics

    Parry, David; Breton, Alice; Abernethy, David; Cockcroft, Sophie; Gillies, John (1999-12)

    Working or discussion paper
    University of Otago

    Since July 1998 we have been teaching an Internet-based distance learning course in health informatics. The development of this course and the experiences we have had running it are described in this paper. The course was delivered using paper materials, a face-to-face workshop, a CD-ROM and Internet communication tools. We currently have about 30 students around New Zealand, a mixture of physicians, nurses and other health staff. Some teaching methods have worked, some haven't, but in the process we have learned a number of valuable lessons.

    View record details
  • Predictive modelling of plankton dynamics in freshwater lakes using genetic programming

    Whigham, Peter A; Recknagel, Friedrich (1999-09)

    Working or discussion paper
    University of Otago

    Building predictive time series models for freshwater systems is important both for understanding the dynamics of these natural systems and in the development of decision support and management software. This work describes the application of a machine learning technique, namely genetic programming (GP), to the prediction of chlorophyll-a. The system endeavoured to evolve several mathematical time series equations, based on limnological and climate variables, which could predict the dynamics of chlorophyll-a on unseen data. The predictive accuracy of the genetic programming approach was compared with an artificial neural network and a deterministic algal growth model. The GP system evolved some solutions which were improvements over the neural network and showed that the transparent nature of the solutions may allow inferences about underlying processes to be made. This work demonstrates that non-linear processes in natural systems may be successfully modelled through the use of machine learning techniques. Further, it shows that genetic programming may be used as a tool for exploring the driving processes underlying freshwater system dynamics.

    View record details
  • Fuzzy logic for software metric models throughout the development life-cycle

    Gray, Andrew; MacDonell, Stephen (1999-09)

    Working or discussion paper
    University of Otago

    One problem faced by managers who are using project management models is the elicitation of numerical inputs. Obtaining these with any degree of confidence early in a project is not always feasible. Related to this difficulty is the risk of precisely specified outputs from models leading to overcommitment. These problems can be seen as the collective failure of software measurements to represent the inherent uncertainties in managers' knowledge of the development products, resources, and processes. It is proposed that fuzzy logic techniques can help to overcome some of these difficulties by representing the imprecision in inputs and outputs, as well as providing a more expert-knowledge based approach to model building. The use of fuzzy logic for project management however should not be the same throughout the development life cycle. Different levels of available information and desired precision suggest that it can be used differently depending on the current phase, although a single model can be used for consistency.

    View record details
  • The development of an electronic distance learning course in health informatics

    Parry, David; Cockcroft, Sophie; Breton, Alice; Abernethy, David; Gillies, John (1999-12)

    Working or discussion paper
    University of Otago

    Since 1997 the authors have been involved in the development of a distance learning course in health informatics. The course is delivered via CD-ROM and the Internet. During this process we have learned valuable lessons about computer-assisted collaboration and cooperative work. In particular we have developed methods of using the software tools available for communication and education. We believe that electronic distance learning offers a realistic means of providing education in health informatics and other fields to students whom for reasons of geography or work commitments would not be able to participate in a conventional course.

    View record details
  • Using rough sets to study expert behaviour in induction of labour

    Parry, David; Yeap, Wai Kiang; Pattison, Neil (1999-12)

    Working or discussion paper
    University of Otago

    The rate of induction of labour (IOL) is increasing, despite no obvious increase in the incidence of the major indications. However the rate varies widely between different centres and practitioners and this does not seem to be due to variations in patient populations. The IOL decision-making process of six clinicians was recorded and examined using hypothetical scenarios presented on a computer. Several rules were identified from a rough sets analysis of the data. These rules were compared to the actual practise of these clinicians in 1994 Initial tests of these rules show that they may form a suitable set for developing an expert system for the induction of labour.

    View record details
  • The NZDIS project: An agent-based distributed information systems architecture

    Purvis, Martin; Cranefield, Stephen; Bush, Geoff; Carter, Dan; McKinlay, Bryce; Nowostawski, Mariusz; Ward, Roy (1999-08)

    Working or discussion paper
    University of Otago

    This paper describes an architecture for building distributed information systems from existing information resources, based on distributed object and software agent technologies. This architecture is being developed as part of the New Zealand Distributed Information Systems (NZDIS) project. An agent-based architecture is used: information sources are encapsulated as information agents that accept messages in an agent communication language (the FIPA ACL). A user agent assists users to browse ontologies appropriate to their domain of interest and to construct queries based on terms from one or more ontologies. One or more query processing agents are then responsible for discovering (from a resource broker agent) which data source agents are relevant to the query, decomposing the query into subqueries suitable for those agents (including the translation of the query into the specific ontologies implemented by those agents), executing the subqueries and translating and combining the subquery results into the desired result set. Novel features of this system include the use of standards from the object-oriented community such as the Common Object Request Broker Architecture (CORBA) (as a communications infrastructure), the Unified Modeling Language (used as an ontology representation language), the Object Data Management Group's Object Query Language (used for queries) and the Object Management Group's Meta Object Facility (used as the basis for an ontology repository agent). Query results need not be returned within an ACL message, but may instead be represented by a CORBA object reference which may be used to obtain the result set.

    View record details