Browsing by Subject "machine learning"

Now showing 1 - 12 of 12
  • Thumbnail Image
    Item
    Open Access
    Advanced analytics for process analysis of turbine plant and components
    (2019) Maharajh,Yashveer; Rousseau, Pieter; Mishra, Amit
    This research investigates the use of an alternate means of modelling the performance of a train of feed water heaters in a steam cycle power plant, using machine learning. The goal of this study was to use a simple artificial neural network (ANN) to predict the behaviour of the plant system, specifically the inlet bled steam (BS) mass flow rate and the outlet water temperature of each feedwater heater. The output of the model was validated through the use of a thermofluid engineering model built for the same plant. Another goal was to assess the ability of both the thermofluid model and ANN model to predict plant behaviour under out of normal operating circumstances. The thermofluid engineering model was built on FLOWNEX® SE using existing custom components for the various heat exchangers. The model was then tuned to current plant conditions by catering for plant degradation and maintenance effects. The artificial neural network was of a multi-layer perceptron (MLP) type, using the rectified linear unit (ReLU) activation function, mean squared error (MSE) loss function and adaptive moments (Adam) optimiser. It was constructed using Python programming language. The ANN model was trained using the same data as the FLOWNEX® SE model. Multiple architectures were tested resulting in the optimum model having two layers, 200 nodes or neurons in each layer with a batch size of 500, running over 100 epochs. This configuration attained a training accuracy of 0.9975 and validation accuracy of 0.9975. When used on a test set and to predict plant performance, it achieved a MSE of 0.23 and 0.45 respectively. Under normal operating conditions (six cases tested) the ANN model performed better than the FLOWNEX® SE model when compared to actual plant behaviour. Under out of normal conditions (four cases tested), the FLOWNEX SE® model performed better than the ANN. It is evident that the ANN model was unable to capture the “physics” of a heat exchanger or the feed heating process as a result of its poor performance in the out of normal scenarios. Further tuning by way of alternate activation functions and regularisation techniques had little effect on the ANN model performance. The ANN model was able to accurately predict an out of normal case only when it was trained to do so. This was achieved by augmenting the original training data with the inputs and results from the FLOWNEX SE® model for the same case. The conclusion drawn from this study is that this type of simple ANN model is able to predict plant performance so long as it is trained for it. The validity of the prediction is highly dependent on the integrity of the training data. Operating outside the range which the model was trained for will result in inaccurate predictions. It is recommended that out of normal scenarios commonly experienced by the plant be synthesised by engineering modelling tools like FLOWNEX® SE to augment the historic plant data. This provides a wider spectrum of training data enabling more generalised and accurate predictions from the ANN model.
  • Thumbnail Image
    Item
    Open Access
    Comparison of sovereign risk and its determinants
    (2019) Smith, Anri; Barr, Graham
    This paper aims to measure, compare and model Sovereign Risk. The risk position of South Africa compared to Emerging Markets as well as in comparison to Developed Markets is considered. Particular interest is taken in how the South African Sovereign Risk environment, and its associated determinants, differs and conforms to that of other Emerging Markets. This effectively highlights how the South African economy is similar to the Emerging Markets and where it behaves differently. Regression, optimisation techniques, dimension reduction techniques as well as Machine Learning techniques, through the use of sentiment analysis, is utilised in this research.
  • Thumbnail Image
    Item
    Open Access
    Computational Psychiatry - Neuropsychological Bayesian reinforcement learning
    (2022) Wolpe, Zach; Shock, Jonathan; Cowley, Benjamin; Clark, Allan
    Cognitive science draws inspiration from a myriad of disciplines, and has become increasingly reliant on computational methods. In particular, theories of learning, operant conditioning and decision making have shown a natural synergy with statistical learning algorithms. This offers a unique opportunity to derive novel insight into the conditioning process by leveraging computational ideas. Specifically, ideas from Bayesian Inference and Reinforcement Learning. In this thesis, we examine the statistical properties of associative learning under uncertainty. We conducted a neuropsychological experiment on over 100 human subjects to measure a suite of executive functions. The primary experimental task (Card Sorting) gauges one's ability to learn, via inference, the structure of some latent pattern that drives the decision making process. We were able to successfully predict the subjects' behaviour in this task by fitting a Bayesian Reinforcement Learning model, alluding to the mechanics of the latent biological decision generating process and executive functions. Primarily, we detail the relationship between working memory capacity and associative learning.
  • Thumbnail Image
    Item
    Open Access
    Computer Vision-Based Classification of Flow Regime and Vapor Quality in Vertical Two-Phase Flow
    (Multidisciplinary Digital Publishing Institute, 2022-01-27) Kadish, Shai; Schmid, David; Son, Jarryd; Boje, Edward
    This paper presents a method to classify flow regime and vapor quality in vertical two-phase (vapor-liquid) flow, using a video of the flow as the input; this represents the first high-performing and entirely camera image-based method for the classification of a vertical flow regime (which is effective across a wide range of regimes) and the first image-based tool for estimating vapor quality. The approach makes use of computer vision techniques and deep learning to train a convolutional neural network (CNN), which is used for individual frame classification and image feature extraction, and a deep long short-term memory (LSTM) network, used to capture temporal information present in a sequence of image feature sets and to make a final vapor quality or flow regime classification. This novel architecture for two-phase flow studies achieves accurate flow regime and vapor quality classifications in a practical application to two-phase CO2 flow in vertical tubes, based on offline data and an online prototype implementation, developed as a proof of concept for the use of these models within a feedback control loop. The use of automatically selected image features, produced by a CNN architecture in three distinct tasks comprising flow-image classification, flow-regime classification, and vapor quality prediction, confirms that these features are robust and useful, and offer a viable alternative to manually extracting image features for image-based flow studies. The successful application of the LSTM network reveals the significance of temporal information for image-based studies of two-phase flow.
  • Thumbnail Image
    Item
    Open Access
    COVID-19 Diagnosis: A Review of Rapid Antigen, RT-PCR and Artificial Intelligence Methods
    (Multidisciplinary Digital Publishing Institute, 2022-04-03) Aruleba, Raphael Taiwo; Adekiya, Tayo Alex; Ayawei, Nimibofa; Obaido, George; Aruleba, Kehinde; Mienye, Ibomoiye Domor; Aruleba, Idowu; Ogbuokiri, Blessing
    As of 27 December 2021, SARS-CoV-2 has infected over 278 million persons and caused 5.3 million deaths. Since the outbreak of COVID-19, different methods, from medical to artificial intelligence, have been used for its detection, diagnosis, and surveillance. Meanwhile, fast and efficient point-of-care (POC) testing and self-testing kits have become necessary in the fight against COVID-19 and to assist healthcare personnel and governments curb the spread of the virus. This paper presents a review of the various types of COVID-19 detection methods, diagnostic technologies, and surveillance approaches that have been used or proposed. The review provided in this article should be beneficial to researchers in this field and health policymakers at large.
  • No Thumbnail Available
    Item
    Open Access
    Detecting and predicting land use and land cover change in the cross-sanaga-bioko coastal forest region for sustainable forest management
    (2024) Njume, Epie Wesner; Hull, Simon
    This study assesses forest, agriculture and built-up areas change in the Cross-Sanaka-Bioko (CSB) region from 2000 to 2021, aiming to provide reliable data for sustainable forest management practices. This analysis will be accomplished with the aid of GIS tools (Google Earth Engine and ArcGIS Pro) and remote sensing data (LULC maps and digital elevation models) in the CSB region. Land use and land cover (LULC) changes in forested regions are critical indicators of environmental transformation, contributing to deforestation, forest degradation, and biodiversity loss, with significant impacts on the environment and human well-being. Sustainable forest management is essential for maintaining ecological balance and ensuring forest resources for future generations. A supervised LULC classification map was created for 2000, 2007, 2014, and 2021 using a decision tree-based machine learning algorithm. Loss, gain and post-classification change detection analysis were used to pinpoint significant LULC changes in the region. Identifying the potential impacts of LULC changes to the environment, air pollutants (CO, NO2, SO2, and PM2.5) were used to first evaluate the variation of emission of the pollutants over the years using a descriptive statistic. Furthermore, a point biserial correlation analysis was used to test the strength of association between the supervised LULC classes with the identified pollutants. Lastly the Multi-Layer Perceptron and Cellular Automata-Markov chain models were used to predict land cover change in the region in the year 2063 and validated by comparing the predicted 2063 map with the 2000 and 2021 classified maps in the CSB region. The study revealed a significant reduction in forested areas (35.55% loss), with the most substantial decline (14.69%) between 2007 and 2014. Agricultural and built-up areas increased by 28.05% and 13.73%, respectively. The primary LULC transition was from forests to agricultural areas, followed by built-up areas. Pollutant emissions, except for NO2, exceeded WHO-recommended values in the region. The results from the correlation analysis showed positive and negative correlations between the LULC changes and air pollutants. For example, agriculture had a moderate positive correlation with NO2 and a moderate negative correlation with CO. There is a projected 21.03% loss in forested areas by 2063, with agricultural lands expanding by 19.69% and built-up areas by 10.88%. These findings highlight the urgent need for sustainable development practices to balance forest conservation, agricultural growth, and urban expansion, aligning with Goal 7 of the African Union Agenda to promote environmental sustainability, and Goal 15, Target 15.2 of the United Nations Sustainable Development Goals.
  • Thumbnail Image
    Item
    Restricted
    How well can post-traumatic stress disorder be predicted from pre-trauma risk factors? an exploratory study in the WHO World Mental Health Surveys
    (2014) Kessler, Ronald C; Rose, Sherri; Koenen, Karestan C; Karam, Elie G; Stang, Paul E; Stein, Dan J; Heeringa, Steven G; Hill, Eric D; Liberzon, Israel; McLaughlin, Katie A; McLean, Samuel A; Pennell, Beth E; Petukhova, Maria; Rosellini, Anthony J; Ruscio, Ayelet M; Shahly, Victoria; Shalev, Arieh Y; Silove, Derrick; Zaslavsky, Alan M; Angermeyer, Matthias C; Bromet, Evelyn J; de Almeida, José Miguel Caldas; de Girolamo, Giovanni; de Jonge, Peter; Demyttenaere, Koen; Florescu, Silvia E; Gureje, Oye; Haro, Josep Maria; Hinkov, Hristo; Kawakami, Norito; Kovess-Masfety, Viviane; Lee, Sing; Medina-Mora, Maria Elena; Murphy, Samuel D; Navarro-Mateu, Fernando; Piazza, Marina; Posada-Villa, Jose; Scott, Kate; Torres, Yolanda; Viana, Maria Carmen
    Post-traumatic stress disorder (PTSD) should be one of the most preventable mental disorders, since many people exposed to traumatic experiences (TEs) could be targeted in first response settings in the immediate aftermath of exposure for preventive intervention. However, these interventions are costly and the proportion of TE-exposed people who develop PTSD is small. To be cost-effective, risk prediction rules are needed to target high-risk people in the immediate aftermath of a TE. Although a number of studies have been carried out to examine prospective predictors of PTSD among people recently exposed to TEs, most were either small or focused on a narrow sample, making it unclear how well PTSD can be predicted in the total population of people exposed to TEs. The current report investigates this issue in a large sample based on the World Health Organization (WHO)'s World Mental Health Surveys. Retrospective reports were obtained on the predictors of PTSD associated with 47,466 TE exposures in representative community surveys carried out in 24 countries. Machine learning methods (random forests, penalized regression, super learner) were used to develop a model predicting PTSD from information about TE type, socio-demographics, and prior histories of cumulative TE exposure and DSM-IV disorders. DSM-IV PTSD prevalence was 4.0% across the 47,466 TE exposures. 95.6% of these PTSD cases were associated with the 10.0% of exposures (i.e., 4,747) classified by machine learning algorithm as having highest predicted PTSD risk. The 47,466 exposures were divided into 20 ventiles (20 groups of equal size) ranked by predicted PTSD risk. PTSD occurred after 56.3% of the TEs in the highest-risk ventile, 20.0% of the TEs in the second highest ventile, and 0.0-1.3% of the TEs in the 18 remaining ventiles. These patterns of differential risk were quite stable across demographic-geographic sub-samples. These results demonstrate that a sensitive risk algorithm can be created using data collected in the immediate aftermath of TE exposure to target people at highest risk of PTSD. However, validation of the algorithm is needed in prospective samples, and additional work is warranted to refine the algorithm both in terms of determining a minimum required predictor set and developing a practical administration and scoring protocol that can be used in routine clinical practice.
  • Thumbnail Image
    Item
    Open Access
    Leveraging big data resources and data integration in biology: applying computational systems analyses and machine learning to gain insights into the biology of cancers
    (2020) Sinkala, Musalula; Martin, Darren; Mulder, Nicola; Barth, Stefan
    Recently, many "molecular profiling" projects have yielded vast amounts of genetic, epigenetic, transcription, protein expression, metabolic and drug response data for cancerous tumours, healthy tissues, and cell lines. We aim to facilitate a multi-scale understanding of these high-dimensional biological data and the complexity of the relationships between the different data types taken from human tumours. Further, we intend to identify molecular disease subtypes of various cancers, uncover the subtype-specific drug targets and identify sets of therapeutic molecules that could potentially be used to inhibit these targets. We collected data from over 20 publicly available resources. We then leverage integrative computational systems analyses, network analyses and machine learning, to gain insights into the pathophysiology of pancreatic cancer and 32 other human cancer types. Here, we uncover aberrations in multiple cell signalling and metabolic pathways that implicate regulatory kinases and the Warburg effect as the likely drivers of the distinct molecular signatures of three established pancreatic cancer subtypes. Then, we apply an integrative clustering method to four different types of molecular data to reveal that pancreatic tumours can be segregated into two distinct subtypes. We define sets of proteins, mRNAs, miRNAs and DNA methylation patterns that could serve as biomarkers to accurately differentiate between the two pancreatic cancer subtypes. Then we confirm the biological relevance of the identified biomarkers by showing that these can be used together with pattern-recognition algorithms to infer the drug sensitivity of pancreatic cancer cell lines accurately. Further, we evaluate the alterations of metabolic pathway genes across 32 human cancers. We find that while alterations of metabolic genes are pervasive across all human cancers, the extent of these gene alterations varies between them. Based on these gene alterations, we define two distinct cancer supertypes that tend to be associated with different clinical outcomes and show that these supertypes are likely to respond differently to anticancer drugs. Overall, we show that the time has already arrived where we can leverage available data resources to potentially elicit more precise and personalised cancer therapies that would yield better clinical outcomes at a much lower cost than is currently being achieved.
  • Thumbnail Image
    Item
    Open Access
    Multilevel inverters for renewable energy systems
    (2018) Chiwaridzo, Pride; Barendse, Paul
    Voltage source inverters have become widely used in the last decade primarily due to the fact that the dangers and limitations of relying on fossil fuel based power generation have been seen and the long term effects felt especially with regards to climate change. Policies and targets have been implemented such as from the United Nations climate change conference (COPxx) concerning human activities that contribute to global warming from individual countries. The most effective way of reducing these greenhouse gases is to turn to renewable energy sources such as the solar, wind etc instead of coal. Converters play the crucial role of converting the renewable source dc power to ac single phase or multiphase. The advancement in research in renewable energy sources and energy storage has made it possible to do things more efficiently than ever before. Regular or 2 level inverters are adequate for low power low voltage applications but have drawbacks when being used in high power high voltage applications as switching components have to be rated upwards and also switch between very high potential differences. To lessen the constraints on the switching components and to reduce the filtering requirements, multilevel inverters (MLI's) are preferred over two level voltage source inverters (VSI's). This thesis discusses the implementation of various types of MLI's and compares four different pulse width modulation (pwm) techniques that are often used in MLI under consideration: three, five, seven and nine level inverters. Harmonic content of the output voltage is recorded across a range of modulation indices for each of the three popular topologies in literature. Output from the inverter is filtered using an L only and an LC filter whose design techniques are presented. A generalized prediction algorithm using machine learning techniques to give the value of the expected THD as the modulation index is varied for a specific topology and PWM switching method is proposed in this study. Simulation and experimental results are produced in five level form to verify and validate the proposed algorithm.
  • Thumbnail Image
    Item
    Open Access
    News media, asset prices and capital flows: evidence from a small open economy
    (2017) Sher, Galen; Strugnell, Dave
    Objectives: This work investigates the role for the content of print news media in determining asset prices and capital flows in a small open economy (South Africa). Specifically, it examines how much of the daily variation in stock prices, bond prices, trading volume and capital flows can be explained by phrases in the print news media. Furthermore, this work links such evidence to the existing theoretical and empirical literature. Methods: This work employs natural language processing techniques for counting words and phrases within articles published in national newspapers. Variance decompositions of the resulting word and phrase counts summarise the information extracted from national newspapers in this way. Following previous studies of the United States, least squares regression relates stock returns to single positive or negative 'sentiment' factors. New in this study, support vector regression relates South African stock returns, bond returns and capital flows to the high-dimensional word and phrase counts from national newspapers. Results: I find that domestic asset prices and capital flows between residents and non-residents reflect the content of domestic print news media. In particular, I find that the contents of national newspapers can predict 9 percent of the variation in daily stock returns one day ahead and 7 percent of the variation in the daily excess return of long-term bonds over short-term bonds three days ahead. This predictability in stocks and bonds coincides with predictability of the content of domestic print news media for net equity and debt portfolio capital inflows, suggesting that the domestic print news media affects foreign residents' demand for domestic assets. Moreover, predictability of domestic print news media for near future stock returns is driven by emotive language, suggesting a role for 'sentiment', while such predictability for stock returns further ahead and the premium on long-term bonds is driven by non-emotive language, suggesting a role for other media factors in determining asset prices. These results do not seem to reflect a purely historical phenomenon, finite-sample biases, reverse causality, serial correlation, volatility or day-of-the-week effects. The results support models where foreign agents' short-run beliefs or preferences respond to the content of domestic print news media heterogeneously from those of domestic agents, while becoming more homogeneous in the medium term.
  • No Thumbnail Available
    Item
    Open Access
    Predicting mergers and acquisitions using machine learning
    (2024) Beckenstrater, Gordon; Marais, Patrick
    Mergers and acquisitions (M&As) play a crucial role in the expansion of companies. During a typical M&A deal, the target company is offered a significant premium over their current share price by the acquirer. This results in a material increase in the target company's share price on the announcement of acquisition. Therefore, accurately forecasting M&As, despite the challenge due to their rarity, presents a lucrative opportunity for investors. Traditional statistical forecasting techniques, reliant on fundamental and technical metrics along with a few macroeconomic indicators, often struggle to pick up underlying relationships between features and targets. This study investigates the effectiveness of advanced machine learning techniques, which have found large success in stock price and fraud prediction, in predicting M&As. logistic regression, a popular statistical technique in M&A literature, serves as a baseline. The performance of algorithms such as random forest, LightGBM, long short-term memory networks (LSTM) and the TabTransformer are evaluated against the baseline. A secondary objective is the development of a robust ensemble model for potential use in an investment portfolio. The algorithms were trained on a comprehensive historical dataset with diverse financial indicators. Given the considerable amount of missing values in the dataset, imputation was applied to allow all algorithms to function properly. Feature selection was conducted to remove redundant features, mitigating their impact on validation performance of the models. Data imbalance was addressed with data sampling techniques which proved substantial in improving validation performance. The findings are that all the advanced algorithms surpassed the performance of logistic regression in M&A prediction, signalling a shift from traditional statistical methods to advanced machine learning techniques. LightGBM and the Ensemble model displayed the best performance in M&A prediction. These results also show that an investment portfolio, constructed based on the most confident predictions of the Ensemble model, forms the basis for a profitable investment strategy.
  • Thumbnail Image
    Item
    Open Access
    Principal points, principal curves and principal surfaces
    (2015) Ganey, Raeesa; Lubbe, Sugnet
    The idea of approximating a distribution is a prominent problem in statistics. This dissertation explores the theory of principal points and principal curves as approximation methods to a distribution. Principal points of a distribution have been initially introduced by Flury (1990) who tackled the problem of optimal grouping in multivariate data. In essence, principal points are the theoretical counterparts of cluster means obtained by the k-means algorithm. Principal curves defined by Hastie (1984), are smooth one-dimensional curves that pass through the middle of a p-dimensional data set, providing a nonlinear summary of the data. In this dissertation, details on the usefulness of principal points and principal curves are reviewed. The application of principal points and principal curves are then extended beyond its original purpose to well-known computational methods like Support Vector Machines in machine learning.