Massey Documents by Type

Permanent URI for this communityhttps://mro.massey.ac.nz/handle/10179/294

Browse

Subject: search.filters.subject.Data processing

Search Results

Now showing 1 - 10 of 81
  • Thumbnail Image
    Item
    Deep learning for low-resource machine translation : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Computer Science, School of Mathematical and Computational Sciences, Massey University, Albany, Auckland, New Zealand. EMBARGOED until further notice.
    (Massey University, 2025-09-01) Gao, Yuan
    Machine translation, a key task in natural language processing, aims to automatically translate text from one language to another while preserving semantic integrity. This thesis builds upon existing research and introduces three deep-learning methods to enhance translation performance under low-resource conditions: (i) an effective transfer learning framework that leverages knowledge from high-resource language pairs, (ii) a pre-ordering-aware training method that explicitly utilizes contextualized representations of pre-ordered sentences, and (iii) a data augmentation strategy that expands the training data size. Firstly, we develop a two-step fine-tuning (TSFT) transfer learning framework for low-resource machine translation. Due to the inherent linguistic divergence between languages in parent (high-resource language pairs) and child (low-resource language pairs) translation tasks, the parent model often serves as a suboptimal initialization point for directly fine-tuning the child model. Our TSFT framework addresses this limitation by incorporating a pre-fine-tuning stage that adapts the parent model to the child source language characteristics, improving child model initialization and overall translation quality. Secondly, we propose a training method that enables the model to learn pre-ordering knowledge and encode the word reordering information within the contextualized representation of source sentences. Pre-ordering refers to rearranging source-side words to better align with the target-side word order before translation, which helps mitigate word-order differences between languages. Existing methods typically integrate the information of pre-ordered source sentences at the token level, where each token is assigned a local representation that fails to capture broader contextual dependencies. Moreover, these methods still require pre-ordered sentences during inference, which incur additional inference costs. In contrast, our method enables the model to encode the pre-ordering information in the contextualized representations of source sentences. In addition, our method eliminates the need for pre-ordering sentences at inference time while preserving its benefits in improving translation quality. Thirdly, to address data scarcity in low-resource scenarios, we propose a data augmentation strategy that employs high-quality translation models trained bidirectionally on high-resource language pairs. This strategy generates diverse, high-fidelity pseudo-training data through systematic sentence rephrasing, generating multiple target translations for each source sentence.. The increased diversity on the target side enhances the model's robustness, as demonstrated by significant performance improvements in eight pairs of low-resource languages. Finally, we conduct an empirical study to explore the potential of applying ChatGPT for machine translation. We design a set of translation prompts incorporating various auxiliary information to assist ChatGPT in generating translations. Our findings indicate that, with carefully designed prompts, ChatGPT can achieve results comparable to those of commercial translation systems for high-resource languages. Moreover, this study establishes a foundation for future research, offering insights into prompt engineering strategies for leveraging large language models in machine translation tasks.
  • Thumbnail Image
    Item
    A framework for enhanced decision making in construction organisations based on quality of pipeline information : a thesis submitted in partial fulfilment of the requirements for the degree of Doctor of Philosophy, School of Built Environment, Massey University, New Zealand
    (Massey University, 2025-08-25) Moshood, Taofeeq Durojaye
    The New Zealand government actively pursues the development of a sustainable construction sector that ensures high performance, productivity, innovation, and community well-being through diverse project initiatives spanning residential, non-residential, and infrastructure development. Despite these aspirations, construction projects in New Zealand are frequently delayed and not effectively completed, thus leading to underperformance. Two critical factors contributing to project delays and suboptimal outcomes are inadequate quality of information and ineffective strategic decision-making processes. Formulating effective strategic decisions is a fundamental challenge for construction organisations, significantly impacting their overall strategic goals and operational success. While information management and decision execution are widely recognised as crucial elements in organisational strategy, there remains a notable gap in understanding the intricate relationship between the quality of information and strategic decision-making, particularly within the context of construction business performance. This research addresses this critical knowledge gap by identifying the determinants for successful construction business performance. It investigates the mediating role of quality of information in the relationship between strategic decision-making and the performance of construction businesses in New Zealand. The study employs a comprehensive methodological approach combining systematic literature review, quantitative analysis, and stakeholder validation to develop a robust understanding of these complex relationships. The systematic literature review focused specifically on identifying critical determinants for successful construction business performance in New Zealand, utilising the ATLAS.ti 9 tool for analysis. This comprehensive review highlighted the fundamental role of the quality of information in strategic decision-making processes and its subsequent impact on organisational performance. This systematic analysis led to the development of a conceptual framework and associated hypotheses establishing the relationships between these factors and their impact on strategic decision-making, information quality, and construction business performance. A quantitative survey was conducted with 102 respondents, and the sample size was determined using G*Power analysis to ensure statistical validity. The study examined how strategic decision-making and information quality influence construction business performance in New Zealand. The research method that was employed integrated established theoretical models and employed Partial Least Squares Structural Equation Modelling (PLS-SEM) using SmartPLS4 software. This analytical approach allowed for robust testing of both direct relationships between variables and the mediating effect of information quality on the relationship between strategic management practices and business performance outcomes. The findings revealed a strong positive correlation between strategic decision-making management and construction business performance, with quality of information serving as a crucial mediating factor. A key research component involved validating the proposed strategic decision-making process framework for its sufficiency, clarity, coherence, relevance, and applicability in New Zealand construction organisations. This validation process included in-depth interviews with six industry experts representing various organisational approaches to strategic decision-making. Their responses revealed a spectrum of techniques ranging from highly structured to more informal methodologies, providing valuable insights into the practical application of strategic decision-making frameworks in different organisational contexts. The research outcomes significantly contribute to theoretical understanding and practical application in the construction industry. Establishing clear criteria for successful quality of information and identifying critical factors affecting project implementation success, this study offers valuable guidance for improving practices within the construction sector. The findings emphasise the crucial importance of prioritising both the quality of information and strategic decision-making to effectively lessen project delays and optimise overall performance in the construction sector. Furthermore, the research contributes to developing more effective strategic decision-making processes by highlighting the interconnected nature of the quality of information and strategic planning. The validated framework provides construction organisations with a practical tool for enhancing their strategic decision-making capabilities while considering the critical role of quality of information in achieving successful outcomes. This study's findings have significant implications for construction industry stakeholders, such as project managers, organisational leaders, and policymakers. The research demonstrates that improving the quality of information and strategic decision-making processes can enhance project outcomes and overall organisational performance. The validated framework offers a structured approach for organisations to assess and improve their current practices while considering their unique operational contexts and requirements.
  • Thumbnail Image
    Item
    Essays on finance and deep learning : a thesis presented in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Finance, School of Economics and Finance, Massey University
    (Massey University, 2025-07-25) Pan, Guoyao
    This thesis aims to broaden the application of deep learning techniques in financial research and comprises three essays that make meaningful contributions to the related literature. Essay One integrates deep learning into the Hub Strategy, a novel chart pattern analysis method, to develop trading strategies. Utilizing deep learning models, which analyze chart patterns alongside data such as trading volume, price volatility, and sentiment indicators, the strategy forecasts stock price movements. Tests on U.S. S&P 500 index stocks indicate that Hub Strategy trading methods, when integrated with deep learning models, achieve an annualized average return of approximately 25%, significantly outperforming the benchmark buy-and-hold strategy's 9.6% return. Risk-adjusted metrics, including Sharpe ratios and Jensen’s alpha, consistently demonstrate the superiority of these trading strategies over both the buy-and-hold approach and standalone Hub Strategy trading rules. To address data snooping concerns, multiple tests validate profitability, and an asset pricing model with 153 risk factors and Lasso-OLS (Ordinary Least Squares) regressions confirms its ability to capture positive alphas. Essay Two utilizes deep learning techniques to explore the relationships between the abnormal return and its explanatory variables, including firm-specific characteristics and realized stock returns. Trained deep learning models effectively predict the estimated abnormal return directly. We evaluate the effectiveness of detecting abnormal returns by comparing our deep learning models against three benchmark methods. When applied to a random dataset, deep learning models demonstrate a significant improvement in identifying abnormal returns within the induced range of -3% to 3%. Moreover, their performance remains consistent across non-random datasets classified by firm size and market conditions. In addition, a regression of abnormal return prediction errors on firm-based factors, market conditions, and periods reveals that deep learning models are less sensitive to variables like firm size, market conditions, and periods than the benchmarks. Essay Three assesses the performance of deep learning predictors in forecasting momentum turning points using the confusion matrix and comparing them to the benchmark model proposed by Goulding, Harvey, and Mazzoleni (2023). Tested on U.S. stocks from January 1990 to December 2023, deep learning predictors demonstrate higher accuracy in identifying turning points than the benchmark. Furthermore, our deep learning-based trading rules yield higher mean log returns and Sharpe ratios, along with lower volatility, compared to the benchmark. Two models achieve average monthly returns of 0.0148 and 0.0177, surpassing the benchmark’s 0.0108. These gains are both economically and statistically significant, with consistent annual results. Regression analysis also shows that our models respond more effectively to changes in stock and market return volatility than the benchmark. Overall, these essays expand the application of deep learning in finance research, demonstrating high predictive accuracy, enhanced trading profitability, and effective detection of long-term abnormal returns, all of which hold significant practical value.
  • Thumbnail Image
    Item
    Fruit measurement horticultural device : developing trust through usability across complex systems : a thesis submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy, College of Creative Arts / Toi Rauwhārangi, Massey University, Wellington, New Zealand
    (Massey University, 2024-12-11) Krige, Zené
    The agricultural technology (ag-tech) sector aims to use emerging technologies to meet changing consumer demands. To do this, the design of an intuitive smart object needed to be developed, and appraised for the horticultural industry of New Zealand. Its subsequent data needed to be expressed in tangible ways that empower decision-making about orchard operations. An elevated user experience of the device, along with quality data driving the system, would provide a successful engagement with an intelligent product system that sustains trust in the interaction and purpose of the product and integrates trust as a value within the system to advance resilience in horticultural innovation. Focusing on the task of fruit measurement, this project explores the conceptual design of a technology-driven device that can efficiently measure fruit size and count, throughout the season. The translation of this data in a format that enables stakeholders to analyse, query and act on it, seeks to inform and empower decision-making by the end users and stakeholders about the best time to harvest. This allows for better management of resources and deployment of labour and equipment. The consequence is a more sustainable orchard operation with greater productivity and benefits to all stakeholders. The project investigates the interrelationships between stakeholders, their equipment and orchard systems to drive product innovation by strengthening foundations of trust and utility, developing confidence in product use, and demonstrating its role in providing critical data into a horticultural management system with an inanimate object (product) placed within the orchard environment. This creative practice research project aims to address the opportunities that design can offer in bridging technological capability to usable products that can communicate trustworthy data clearly to end-users.
  • Thumbnail Image
    Item
    The multimodality of creaminess perception : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Food Technology at Massey University, Manawatū Campus, Palmerston North, New Zealand. EMBARGOED until 21 August 2026.
    (Massey University, 2024-02-28) Fisher, Emily Claire
    Creaminess is a complex sensory sensation that drives consumer acceptability of milk. To date, creaminess research has focused on instrumental and compositional measures overlooking the critical consumer perspective. This research took a consumer-led approach to unlock new insights into the underlying sensory attributes driving consumer creaminess perception using perceptual modelling. Robust sensory data, from a trained panel, was combined with consumer approaches for accurate modelling. Initially, attributes and modalities perceived to drive milk creaminess were identified through discussion with consumers representative of two key dairy markets, China and New Zealand (NZ). Subsequently, a milk sample set (n=32) was developed, and an expert panel trained to profile the samples based on attributes identified by consumers. A novel methodological investigation, on the impact of panel training with Polarised Sensory Positioning (PSP) of the sample set, was also explored. Focusing on NZ consumers, participants (n=117) evaluated creaminess and liking perception of the milk samples. Critically, regression modelling was employed to identify key attributes driving creaminess perception based on expert panel data. Several novel findings were discovered. Drivers of creaminess differed to some degree between NZ and Chinese consumers indicating cultural differences across markets. Trained panel sensory data revealed multicollinearity between attributes measured to describe the sample set. Modelling approaches were able to identify key attributes required to predict creaminess. New findings that training has little impact on PSP outcomes was also ascertained. Pertinently combining four attributes, across different modalities, in an Elastic net regression model (‘yellow’, ‘watery’ flavour, ‘in-mouth thickness’ and ‘astringency’) successfully predicted creaminess (R2=0.9514), however these attributes were highly correlated with others retained in a PLS model. Each model had its relative merits. Of further note, consumer creaminess response was highly variable and cluster analysis revealed two different consumer segments with perception impacted by sensitivity to certain attributes: ‘green tinge’, ‘cardboard’, ‘salty’, ‘cooked’, ‘fat separation’, ‘grassy’, ‘buttery’, ‘melting’, ‘cream’ aroma, ‘smoothness’, and ‘astringent’. This research revealed new understanding concerning perceptual attributes contributing to consumer creaminess perception and provided clearer targets for the dairy industry to ensure milk creaminess levels align to consumer expectations and related commercial gain.
  • Thumbnail Image
    Item
    Explainable spectral super-resolution based on a single RGB image : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy (PhD) in Electronics and Computer Engineering at Massey University, Manawatu, New Zealand
    (Massey University, 2024-04-09) Chang, Yuan
    Hyperspectral imaging offers fine spectral measurements of target surfaces, finding utility in various fields. However, traditional hyperspectral systems grapple with high-cost issues. On the other hand, conventional RGB cameras, which provide relatively coarse measurements of surface spectra, are widely accessible. Consequently, the recovery of spectral information from RGB images has emerged as a popular approach for low-cost hyperspectral imaging, a venture also known as single-image spectral super-resolution. Yet, existing methods, mostly rooted in deep convolutional neural networks, tend to suffer from limited interpretability. In our research, we propose an explainable method for single-image spectral super-resolution. This method relies on the RGBPQR colour space, a low-dimensional spectral data model representing the spectrum. Leveraging the RGBPQR spectral model, we can transform the spectral reconstruction task into a regression problem. To tackle the metamerism issue, we analysed existing spectral super-resolution networks and discovered that these networks often depend on local textural information as context to mitigate metamerism. Informed by this insight, we utilized features extracted from multiscale local binary patterns as contextual information to design our explainable method. Furthermore, in this study, we discussed the error measurements and loss functions employed in this research area and proposed a new error measurement that can represent performance more accurately. We also endeavoured to put forward a method for quantitatively measuring the ability to resolve metamerism, a critical problem in spectral super-resolution. Through our research, we offered a simple, low-dimensional, and explainable spectral super-resolution solution.
  • Thumbnail Image
    Item
    Analysing underpinning patterns in social media posts that promote fat stigmatisation : a thesis submitted in partial fulfilment of the requirements for the degree of Doctor of Philosophy (Ph.D) in Information Technology, Massey University, Auckland, New Zealand
    (Massey University, 2022) Wanniarachchi, Vajisha Udayangi
    Social media offers users an online platform to freely express themselves; however, when users post opinionated and offensive comments that target certain communities, this could instigate hatred towards them. With the global increase in obese/fat populations, social media discourses laced with fat hatred have become commonplace, leading to much fat stigmatising content being posted online. This research aims to investigate the patterns of fat stigma, and how female and male genders are positioned in fat stigmatising discourses that are being conducted over social media. To achieve this objective, a methodological framework is proposed for unearthing underlying stigmatising patterns prevalent in social media discussions, with specific focus on fat stigma. Methods incorporating natural language processing techniques such as sentiment analysis and topic modelling, along with discourse analysis have been described for classifying users’ emotions and comprehending the stigma patterns embedded in social big data. The framework has been applied to weight-based textual data, extracted from Twitter and Reddit, to identify emergent gender-based themes, emotions and word frequency patterns that underpin the fat stigmatising content posted online. The experiential consequences of being considered fat across both genders have been analysed with objectification theory. The findings from this study have provided a holistic outlook on fat stigmatising content that is posted online which can further inform policymakers in planning suitable props to facilitate more inclusive social media spaces. This study showcases how lexical analytics can be conducted by combining a variety of data mining methods to draw out insightful subject related themes that add to the existing knowledge base; therefore, has both practical and theoretical implications.
  • Thumbnail Image
    Item
    Representation learning for the graph data : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Computer Science, Massey University, Albany, Auckland, New Zealand
    (Massey University, 2022) Gan, Jiangzhang
    Graph data consist of the association information between complex entities and also contain diverse vertex information. To make graph data analysis simple and effective, as the bridge between the original graph data and the graph application tasks, graph representation learning has become a hot research topic in recent years. Previous representation learning methods for the graph data may not reflect the intrinsic relationship between nodes due to the complexity of the graph data. Moreover, they do not preserve the topology of the graph data well, which will affect the effectiveness of the downstream tasks. To deal with these issues, the thesis studies effective graph representation learning methods in terms of graph construction and representation learning. We propose a traditional graph learning method under semi-supervised learning to explore parameter-free fusion of graph learning. Specifically, we first employ the Pearson correlation coefficient to obtain a fully connected Functional Connectivity brain Networks (FCN), and then to learn a sparsely connected FCN for every subject. Finally, the ℓ1-SVM is employed to learn the important features and conduct disease diagnosis. We propose an end-to-end deep graph learning method under semi-supervised learning to improve the quality of initial graph. Specifically, the proposed method first extracts the common information and the complementary information among multiple local graphs to obtain a unified local graph, which is then fused with the global graph of the data to obtain the initial graph for the GCN model. As a result, the proposed method conducts the graph fusion process twice to simultaneously learn the low-dimensional space and the intrinsic graph structure of the data in a unified framework. We propose a multi-view unsupervised graph learning method. Specifically, the adaptive data augmentation first builds a feature graph from the feature space, and then designs a deep graph learning model on the original representation and the topology graph, respectively, to update the feature graph and the new representation. As a result, the adaptive data augmentation outputs multi-view information, which is fed into two GCNs to generate multi-view embedding features. Two kinds of contrastive losses are further designed on multi-view embedding features to explore the complementary information among the topology and feature graphs. Additionally, adaptive data augmentation and contrastive learning are embedded in a unified framework to form an end-to-end model. All proposed methods are evaluated on real-world data sets. Experimental results demonstrate that our methods outperformed all comparison methods, compared to state-of-the-art methods.
  • Thumbnail Image
    Item
    Deep learning-based approaches for plant disease and weed detection : a thesis by publications presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Engineering, Massey University, Auckland, New Zealand
    (Massey University, 2022) Saleem, Muhammad Hammad
    To match the ever-growing food demand, the scientific community has been actively focusing on addressing the various challenges faced by the agricultural sector. The major challenges are soil infertility, abrupt changes in climatic conditions, scarcity of water, untrained labor, emission of greenhouse gases, and many others. Moreover, plant diseases and weeds are two of the most important agricultural problems that reduce crop yield. Therefore, accurate detection of plant diseases and weeds is one of the essential operations to apply targeted and timely control measures. As a result, this can improve crop productivity, reduce the environmental effects and financial losses resulting from the excessive application of fungicide/herbicide spray on diseased plants/weeds. Among various ways of plant disease and weed detection, image-based methods are significantly effective for the interpretation of the distinct features. In recent years, image-based deep learning (DL) techniques have been reported in literature for the recognition of weeds and plant diseases. However, the full potential of DL has not yet been explored as most of the methods rely on modifications of the DL models for well-known and readily available datasets. The current studies lack in several ways, such as addressing various complex agricultural conditions, exploring several aspects of DL, and providing a systematic DL-based approach. To address these research gaps, this thesis presents various DL-based methodologies and aims to improve the mean average precision (mAP) for the identification of diseases and weeds in several plant species. The research on plant disease recognition starts with a publicly available dataset called PlantVillage and comparative analyses are conducted on various DL feature extractors, meta-architectures, and optimization algorithms. Later, new datasets are generated from various local New Zealand horticultural farms, named NZDLPlantDisease-v1 & v2. The proposed datasets consist of healthy and diseased plant organs of 13 economically important horticultural crops of New Zealand, divided into 48 classes. A performance-optimized DL model and a transfer learning-based approach are proposed for the detection of plant diseases using curated datasets. The weed identification has been performed on an open-source dataset called DeepWeeds. A two-step weed detection pipeline is presented to show the performance improvement of the deep learning model with a significant margin. The results for both agricultural tasks achieve superior performance compared to the existing method/default settings. The research outcomes elaborate the practical aspects and extended potential of DL for selected agricultural applications. Therefore, this thesis is a benchmark step for cost-effective crop protection and site-specific weed management systems (SSWM).
  • Thumbnail Image
    Item
    Deep learning for asteroid detection in large astronomical surveys : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Computer Science at Massey University, Albany, New Zealand
    (Massey University, 2022) Cowan, Preeti
    The MOA-II telescope has been operating at the Mt John Observatory since 2004 as part of a Japan/NZ collaboration looking for microlensing events. The telescope has a total field of view of 1.6 x 1.3 degrees and surveys the Galactic Bulge several times each night. This makes it particularly good for observing short duration events. While it has been successful in discovering exoplanets, the full scientific potential of the data has not yet been realised. In particular, numerous known asteroids are hidden amongst the MOA data. These can be clearly seen upon visual inspection of selected images. There are also potentially many undiscovered asteroids captured by the telescope. As yet, no tool exists to effectively mine archival data from large astronomical surveys, such as MOA, for asteroids. The appeal of deep learning is in its ability to learn useful representations from data without significant hand-engineering, making it an excellent tool for asteroid detection. Supervised learning requires labelled datasets, which are also unavailable. The goal of this research is to develop datasets suitable for supervised learning and to apply several CNN-based techniques to identify asteroids in the MOA-II data. Asteroid tracklets can be clearly seen by combining all the observations on a given night and these tracklets form the basis of the dataset. Known asteroids were identified within the composite images, forming the seed dataset for supervised learning. These images were used to train several CNNs to classify images as either containing asteroids or not. The top five networks were then configured as an ensemble that achieved a recall of 97.67%. Next, the YOLO object detector was trained to localise asteroid tracklets, achieving a mean average precision (mAP) of 90.97%. These trained networks will be applied to 16 years of MOA archival data to find both known and unknown asteroids that have been observed by the telescope over the years. The methodologies developed can also be used by other surveys for asteroid recovery and discovery.