Browsing by Author "Greene, John"
Now showing 1 - 5 of 5
Results Per Page
Sort Options
- ItemOpen AccessThe application of neural networks to communication channel equalisation : a comparison between localised and non-localised basis functions(1997) Olshewsky, Avron Bernard; Greene, JohnNeural networks have been applied to a number of problems over the past few years. One of the emerging applications of neural networks is adaptive communication channel equalisation. This area of research has become prominent due to the reformulation of the equalisation problem as a classification problem. Viewing equalisation as a classification problem allows researchers to apply the knowledge gained from other fields to equalisation. A wide variety of neural network structures have been suggested to equalise communication channels. Each structure may in turn have a number of different possible algorithms to train the equaliser. A neural network is essentially a non-linear classifier; in general a neural network is able to classify data by employing a non-linear function. The primary subject of this dissertation is the comparative performance of neural networks employing non-localised basis (non-linear) functions (Multi-layer Perceptron) versus those employing localised basis functions (Radial Basis Function Network).
- ItemOpen AccessEvaluation of two prototype three phase photovoltaic water pumping systems(1994) Scholle, Axel; Cowan, Bill; Greene, JohnTwo prototype three phase AC photovoltaic pump systems (Solvo, ML T) and a DC PV pump (Miltek) were tested on a farm borehole in Namibia (latitude 21°6', longitude 17°6'). The PV array consisted of twelve modules (636Wpeak) mounted on a single-axis passive tracker. The depth of the water was 75m and a progressive cavity pump with a self-compensating stator was used in all the tests. Customised data acquisition was designed to measure performance characteristics through a range of operating conditions (mainly steady state); a secondary data acquisition system was used to capture samples of high frequency signals. The data allowed detailed analysis of system, subsystem and component performance, as well as performance evaluation over Standard Solar Days. The focus of the investigation was evaluation of the AC prototypes, in terms of performance, other technical factors, reliability and economic criteria. The analog-based DC system served as a basis for comparison. Both AC systems employed microprocessor control and PWM variable-frequency variable-voltage inversion. Efficiencies, optimality, stability, start-up behaviour, non-productive operating modes and protection were examined. A number of recommendations were proposed for improvements in the basic control algorithms, monitoring and managing non-productive modes, improved protection, layout and user diagnostic features.
- ItemOpen AccessAn exploration into the sparse representation of spectra(2007) Mthembu, Linda; Greene, JohnThis thesis describes an exploration in achieving sparse representations of object, with special focus on spectral data. Given a database of objects one would like to know the actual aspects of each class that distinguish it from any other class in the database. We explore the hypothesis that simple abstractions (descriptions) that humans normally make, especially based on the visual phenomenology or physics on the problem, can be helpful in extracting and formulating useful sparse representations of the observed objects. In this thesis we focus on the discovery of such underlying features, employing a number of recent methods from machine learning. Firstly we find that an approach to automatic feature discovery recently proposed in the literature (Non Negative Matrix Factorization) is not as it seems. We show the limitations of this approach and demonstrate a more efficient method on a synthetic problem. Secondly we explore a more empirical approach to extracting visually attractive features of spectra from which we formulate simple re-representation of spectral data and show that the identification and discovery of certain intuitive features at various scales can be sufficient to describe a spectrum profile. Finally we explore a more traditional and principled automatic method of analyzing a spectrum at different resolutions (Wavelets). We find that certain classes of spectra can easily be discriminated between by a simple approximation of the spectrum profile while in other cases only the finer profile details are important. Throughout this thesis we employ a measure called the separability index as our measure of how easy it is to discriminate objects in a database with the proposed representations.
- ItemOpen AccessMachine learning for corporate failure prediction : an empirical study of South African companies(2004) Kornik, Saul; Everingham, Geoff; Greene, JohnThe research objective of this study was to construct an empirical model for the prediction of corporate failure in South Africa through the application of machine learning techniques using information generally available to investors. The study began with a thorough review of the corporate failure literature, breaking the process of prediction model construction into the following steps: * Defining corporate failure * Sample selection * Feature selection * Data pre-processing * Feature Subset Selection * Classifier construction * Model evaluation These steps were applied to the construction of a model, using a sample of failed companies that were listed on the JSE Securities Exchange between 1 January 1996 and 30 June 2003. A paired sample of non-failed companies was selected. Pairing was performed on the basis of year of failure, industry and asset size (total assets per the company financial statements excluding intangible assets). A minimum of two years and a maximum of three years of financial data were collated for each company. Such data was mainly sourced from BFA McGregor RAID Station, although the BFA McGregor Handbook and JSE Handbook were also consulted for certain data items. A total of 75 financial and non-financial ratios were calculated for each year of data collected for every company in the final sample. Two databases of ratios were created - one for all companies with at least two years of data and another for those companies with three years of data. Missing and undefined data items were rectified before all the ratios were normalised. The set of normalised values was then imported into MatLab Version 6 and input into a Population-Based Incremental Learning (PBIL) algorithm. PBIL was then used to identify those subsets of features that best separated the failed and non-failed data clusters for a one, two and three year forward forecast period. Thornton's Separability Index (SI) was used to evaluate the degree of separation achieved by each feature subset.
- ItemOpen AccessSelf mixing oscillators at Q-band using Gunn diode solid state sources(1982) Braun, Robin Michael; Greene, JohnThis Thesis is a report of research done into the use of a GUNN Diode Millimetre Wave Source as both the Receiver and the Transmitter of a system. The Thesis is divided into three main sections. The first of these sections (Chapter 2) "sets the scene" so to speak. It presents certain preliminary fundamentals that are essential for placing the results obtained later into their proper context. Areas discussed are NOISE FIGURE, TRANSMISSION PATH LOSS and GUNN DIODE operation. The second section (Chapter 3) considers the operation of the device as a mixer. Its operation as a Microwave Power Source is a well-known concept and it will not be discussed here. This section covers such aspects as CONVERSION LOSS, NOISE FIGURE and BIAS CIRCUIT STABILITY. The third section considers those points arising from the second section which seem to be of interest. These include such aspects as Better Correlation of results and theory and ways of applying the device to modern Waveguide structures. The Thesis ends with a Conclusion, Appendix and a Bibliography.