Browsing by Subject "Machine Learning"

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Open Access
    APIC: A method for automated pattern identification and classification
    (2017) Goss, Ryan Gavin; Nitschke, Geoff Stuart
    Machine Learning (ML) is a transformative technology at the forefront of many modern research endeavours. The technology is generating a tremendous amount of attention from researchers and practitioners, providing new approaches to solving complex classification and regression tasks. While concepts such as Deep Learning have existed for many years, the computational power for realising the utility of these algorithms in real-world applications has only recently become available. This dissertation investigated the efficacy of a novel, general method for deploying ML in a variety of complex tasks, where best feature selection, data-set labelling, model definition and training processes were determined automatically. Models were developed in an iterative fashion, evaluated using both training and validation data sets. The proposed method was evaluated using three distinct case studies, describing complex classification tasks often requiring significant input from human experts. The results achieved demonstrate that the proposed method compares with, and often outperforms, less general, comparable methods designed specifically for each task. Feature selection, data-set annotation, model design and training processes were optimised by the method, where less complex, comparatively accurate classifiers with lower dependency on computational power and human expert intervention were produced. In chapter 4, the proposed method demonstrated improved efficacy over comparable systems, automatically identifying and classifying complex application protocols traversing IP networks. In chapter 5, the proposed method was able to discriminate between normal and anomalous traffic, maintaining accuracy in excess of 99%, while reducing false alarms to a mere 0.08%. Finally, in chapter 6, the proposed method discovered more optimal classifiers than those implemented by comparable methods, with classification scores rivalling those achieved by state-of-the-art systems. The findings of this research concluded that developing a fully automated, general method, exhibiting efficacy in a wide variety of complex classification tasks with minimal expert intervention, was possible. The method and various artefacts produced in each case study of this dissertation are thus significant contributions to the field of ML.
  • Thumbnail Image
    Item
    Open Access
    Machine learning for corporate failure prediction : an empirical study of South African companies
    (2004) Kornik, Saul; Everingham, Geoff; Greene, John
    The 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.
  • Thumbnail Image
    Item
    Open Access
    Mobile phone technology as an aid to contemporary transport questions in walkability, in the context of developing countries
    (2019) Chege, Wilberforce Wanjau; Zuidgeest, Marcus
    The emerging global middle class, which is expected to double by 2050 desires more walkable, liveable neighbourhoods, and as distances between work and other amenities increases, cities are becoming less monocentric and becoming more polycentric. African cities could be described as walking cities, based on the number of people that walk to their destinations as opposed to other means of mobility but are often not walkable. Walking is by far the most popular form of transportation in Africa’s rapidly urbanising cities, although it is not often by choice rather a necessity. Facilitating this primary mode, while curbing the growth of less sustainable mobility uses requires special attention for the safety and convenience of walking in view of a Global South context. In this regard, to further promote walking as a sustainable mobility option, there is a need to assess the current state of its supporting infrastructure and begin giving it higher priority, focus and emphasis. Mobile phones have emerged as a useful alternative tool to collect this data and audit the state of walkability in cities. They eliminate the inaccuracies and inefficiencies of human memories because smartphone sensors such as GPS provides information with accuracies within 5m, providing superior accuracy and precision compared to other traditional methods. The data is also spatial in nature, allowing for a range of possible applications and use cases. Traditional inventory approaches in walkability often only revealed the perceived walkability and accessibility for only a subset of journeys. Crowdsourcing the perceived walkability and accessibility of points of interest in African cities could address this, albeit aspects such as ease-of-use and road safety should also be considered. A tool that crowdsources individual pedestrian experiences; availability and state of pedestrian infrastructure and amenities, using state-of-the-art smartphone technology, would over time also result in complete surveys of the walking environment provided such a tool is popular and safe. This research will illustrate how mobile phone applications currently in the market can be improved to offer more functionality that factors in multiple sensory modalities for enhanced visual appeal, ease of use, and aesthetics. The overarching aim of this research is, therefore, to develop the framework for and test a pilot-version mobile phone-based data collection tool that incorporates emerging technologies in collecting data on walkability. This research project will assess the effectiveness of the mobile application and test the technical capabilities of the system to experience how it operates within an existing infrastructure. It will continue to investigate the use of mobile phone technology in the collection of user perceptions of walkability, and the limitations of current transportation-based mobile applications, with the aim of developing an application that is an improvement to current offerings in the market. The prototype application will be tested and later piloted in different locations around the globe. Past studies are primarily focused on the development of transport-based mobile phone applications with basic features and limited functionality. Although limited progress has been made in integrating emerging advanced technologies such as Augmented Reality (AR), Machine Learning (ML), Big Data analytics, amongst others into mobile phone applications; what is missing from these past examples is a comprehensive and structured application in the transportation sphere. In turn, the full research will offer a broader understanding of the iii information gathered from these smart devices, and how that large volume of varied data can be better and more quickly interpreted to discover trends, patterns, and aid in decision making and planning. This research project attempts to fill this gap and also bring new insights, thus promote the research field of transportation data collection audits, with particular emphasis on walkability audits. In this regard, this research seeks to provide insights into how such a tool could be applied in assessing and promoting walkability as a sustainable and equitable mobility option. In order to get policy-makers, analysts, and practitioners in urban transport planning and provision in cities to pay closer attention to making better, more walkable places, appealing to them from an efficiency and business perspective is vital. This crowdsourced data is of great interest to industry practitioners, local governments and research communities as Big Data, and to urban communities and civil society as an input in their advocacy activities. The general findings from the results of this research show clear evidence that transport-based mobile phone applications currently available in the market are increasingly getting outdated and are not keeping up with new and emerging technologies and innovations. It is also evident from the results that mobile smartphones have revolutionised the collection of transport-related information hence the need for new initiatives to help take advantage of this emerging opportunity. The implications of these findings are that more attention needs to be paid to this niche going forward. This research project recommends that more studies, particularly on what technologies and functionalities can realistically be incorporated into mobile phone applications in the near future be done as well as on improving the hardware specifications of mobile phone devices to facilitate and support these emerging technologies whilst keeping the cost of mobile devices as low as possible.