7440 research outputs found
Sort by
Oral narrative intervention with children with autism spectrum disorder and language disorder
No full textThesis by publication.Includes bibliographical references.Chapter 1. Introduction -- Chapter 2. The effects of oral narrative intervention on the narratives of children with language disorder: A systematic literature review -- Chapter 3. The effects of an oral narrative intervention on the personal narratives of children with ASD and severe language disorder -- Chapter 4. The effects of an oral narrative intervention on the fictional narrative retells of children with ASD and severe language impairment: A pilot study -- Chapter 5. The effects of an oral narrative intervention on the original fictional narratives of children with ASD and language disorder -- Chapter 6. Conclusion -- Appendices.Narratives are temporally sequenced accounts of events and have a broad importance to social, academic and language development. Individuals with autism spectrum disorder (ASD) characteristically have difficulties with social communication and have been found to produce narratives that are less structurally complex than typically developing peers. Five papers are included in this thesis by publication. In the first paper a systematic literature review is presented examining research into the effects of narrative intervention on the oral narratives of children with communication disorders. Overall the literature supports the use of narrative intervention to develop narrative macrostructure in children with oral communication disorders. Only four studies included participants with a diagnosis of ASD. Existing research supports the use of macrostructure icons, participants telling entire narratives and clinician modelling within intervention. In the remaining four papers intervention studies are described that address the development of oral narrative macrostructure. In the first intervention study, a multiple baseline with probe design was used to examine the effects of intervention on the personal narratives of four children with ASD and severe language disorder. The same design was used to examine the effects of intervention on the fictional narrative retells on four children with ASD and severe language disorder. The third intervention study was a pilot AB study with one participant with ASD and language disorder that examined an original fictional narrative intervention. Following on from this pilot study, a multiple baseline with probe design was employed in the final intervention study to examine an original fictional narrative intervention with four children with ASD and language disorder. All studies included maintenance and generalisation probes. Overall, the narrative interventions appeared to be effective in improving the macrostructure for most participants, but individual responses were variable and modifications to the intervention procedures were necessary for some participants. There was evidence of maintenance for most participants and some evidence of generalisation to other people and settings.This thesis adds to the limited research into narrative intervention with children with ASD and significant language disorder. The studies presented provide evidence that intervention can be effective across a range of narrative types but individual responses to intervention may be idiosyncratic and some children may require individualised adjustments. The thesis also provides the first study of narrative retell intervention with this population.1 online resource (x, 225 pages
Shining light on the dark Milky Way: probing our galaxy's hidden gas
No full textTheoretical thesis.Bibliography: pages 125-139.1. Introduction -- 2. Properties of warm and cold atomic hydrogen in the Taurus and Gemini regions -- 3. Dust - gas relations and OH abundance -- 4. OH as a reliable tracer of "CO-dark" gas -- 5. Conclusions and future work.The evolution of galaxies is driven by the gas-to-stars cycle of matter in the interstellar medium (ISM). Throughout this process, the transition of hydrogen gas from its atomic (HI) to molecular (H2) form is at the heart of the story. This transition is strongly coupled with the evolution of interstellar dust, as dust grains play a key role in the formation of molecules. However, despite decades of effort, there still remains much uncertainty on the constraints for the physical properties of both the gas and dust in the ISM, due to the limits of our observational capabilities. Presented here in this thesis are attempts to address some of these outstanding issues via observations of atomic and molecular media and detailed data analyses of thermal dust emission and reddening. To begin with, I present a large-scale survey of H I emission and absorption using the Arecibo radio telescope, which explores the physical properties of the cold and warm atomic gas in the vicinity of five giant molecular clouds (GMCs): Taurus, California, Rosette, Mon OB1, NGC 2264. Strong HI absorption was detected towards all 79 background continuum sources in the 60 × 20 square degree area of the sky. Both Gaussian and pseudo-Voigt spectral decompositions were performed to derive the temperatures, optical depths and column densities of the cold and warm atomic gas. The properties of the cold gas in these regions of interest appear to be universal with an excitation (or spin) temperature of ∼ 50 K, consistent with existing estimates. The fraction of warm gas is about 60%, and nearly 40% of this warm gas resides in thermally unstable regime 500-5000 K. There is more cold gas around GMCs than in diffuse regions. The optically thin assumption, on average, underestimates HI column density by ∼ 24%. Comparing the total HI column densities with those derived under the optically thin assumption also allowed me to assess the impact of opacity effects, and determine the most suitable method for opacity correction - that is to apply region-dependent uniform spin temperatures for available HI emission data.Next, I dived into several data sets including opacity-corrected HI column densities from the Arecibo Millennium Survey and 21-SPONGE, together with thermal dust emission data from the Planck satellite, new dust reddening maps from Pan-STARRS 1 and 2MASS, newly published Millennium Survey OH data, and follow-up CO emission observations. In combining all of these datasets, I confirmed linear correlations between dust optical depth τ 353 , dust reddening E ( B − V ) and total gas column density N H in the range (1-30) × 10 20 cm − 2 , along atomic sightlines with no molecular gas detections. I found a 60% higher N H / E ( B − V ) ratio than the current "standard" value, and was able to disentangle the effects of dust grain evolution and so-called "dark gas" (the gas component not detected by H I and CO) on the variation of dust opacity σ 353 = τ 353 / N H . I also derived an OH/H 2 abundance ratio of X OH ∼ 1 × 10 − 7 , finding no evidence for systematic trends in X OH with molecular column density N H 2 in the range N H 2 ∼ ( 1 - 100 ) × 10 20 cm − 2 . These results suggest that the OH molecule may be used as a reliable tracer for the molecular ISM this range, which includes sightlines with both CO-dark and CO-bright gas.Finally, I was involved as co-author in a third project on OH emission-absorption measurements and dark molecular gas. We used OH absorption measurements from the Arecibo Millennium Survey and our own follow-up CO survey. The derived excitation temperatures of the two OH main lines at 1665 and 1667 MHz peak at ∼ 3.5 K, close to the background continuum temperature (CMB + synchrotron), providing an explanation for why OH is often difficult to detect in emission. The OH main lines are optically thin with optical depths always below 0.25; and in general these lines are not in local thermodynamic equilibrium, with a difference in excitation temperatures of | T ex ( 1667 ) − Tex (1665) < 2K. About half of all detected OH absorption components do not have corresponding CO emission detections. This implies that the OH ground-state main lines trace more effectively molecular gas than traditional CO emission. In this work, I per- formed all the OH analysis, which is the main part of this study. However, I did not contribute to the discussions on CO analysis and the dark molecular gas.1 online resource (xviii, 139 pages
Toward a vocational balanced scorecard: a grounded model based on multiple case studies of registered training organizations in Australia
No full textTheoretical thesis.Bibliography: pages 249-286.Chapter 1. Introduction -- Chapter 2. Organisational performance and balanced scorecard -- Chapter 3. Vocational and training sector -- Chapter 4. Research methods and design -- Chapter 5. Analysis and results -- Chapter 6. Discussion and implications -- Chapter 7. Conclusion, limitations and directions for future research.The Balanced Scorecard (BSC) is a widely used management tool to map organizational goals towards execution of strategies in line with the organization's vision. Prior research has either used the concept to develop goals or key performance indicators (KPIs) for different divisions, such as marketing and HR, or focused on the varying levels of successful adoption of BSC in ad hoc case studies. But little has been done to explore if BSC can be applied to a key economic sector where different organizations of fer unique services. While BSC has been investigated across industries, from hospitality to education, there are no studies on the application of BSC at vocational training institutions. The objective of this thesis is to explore the application of BSC for strategic performance in registered training organizations (RTOs), which are the equivalent of vocational training institutions (VTIs) in Australia. Using a sample of 15 RTOs and 30 in - depth interviews with managers, a grounded model for the application of BSC at VTIs called the vocational balanced scorecard (VBSC) was developed. The findings indicate that, as the primary strategic goal of RTOs is to offer higher education courses compatible with universities, the VBSC can significantly help achieve this goal. The VBSC helps set key objectives across course development, staff training, financial management, students' compliancy and support as well as with the process of continuous improvement and quality control. The model developed shows how the selected areas map, according to BSC, and enables RTO managers to adjust and communicate their strategy and rectify procedural deficiencies in their operations.1 online resource (300 pages) illustration
Accounting for greenhouse gas emissions in cities
No full textTheoretical thesis.Bibliography: pages 179-229.Chapter 1: Introduction -- Chapter 2: Cities lead climate Change by mitigating GHG Emission -- Chapter 3: Literature review on social and environmental accounting research -- Chapter 4: Paper 1: greenhouse gas emissions disclosure by cities: the expectation gap -- Chapter 5: Paper 2: measuring for climate actions: a disclosure study of ten megacities -- Chapter 6: Paper 3: energy efficiency public disclosure of major Australian cities -- Chapter 7: Conclusion.Several social and environmental accounting (SEA) scholars have suggested that sustainability issues need to be explored at the regional or geographic boundary level as opposed to focusing on the legal form of corporations (Dumay et al., 2010; Gray, 20 10; Milne and Gray, 2007). One such regional level is that of a city as each city has its own geographic/regional boundary. In relation to the key sustainability issue of climate change, cities are responsible for a large share of anthropogenic greenhouse gas (GHG) emissions. Cities also have the ability and capacity to reduce GHG emissions via strategies and other actions. Policies and actions already being implemented at city level have the potential to reduce GH G emissions by 1 billion metric tons annually by 2030 (Ostrander and Oliveira, 2013). Measurement is vital for managing GHG emissions and disclosure is the critical step for public accountability. Extant accounting literature, however, is relatively silent about exploring the measurement and disclosure practices of GHG emissions at the city level. Prior studies have predominately examined the accounting and reporting of GHG emissions in the corporate sector (e.g., Andrew and Cortese, 2011a; Depoers et al., 2016; Kolk et al., 2008). These studies found that the quality of disclosure is poor and advocated mandatory reporting as well as calling for further GHG disclosure research. In responding to this call, this thesis ex amines the extent and quality of disclosures of GHG emissions and reduction activities at the city-level.This thesis consists of three empirical research papers. Paper 1 (Mia et al., 2019) investigates the quality of GHG emissions disclosures of 42 cities to the Carbon Disclosure Project (CDP) and compares them with the expectations of users. The expectation gap framework is used to examine GHG disclosure quality on the premise that quality disclosure is needed to meet user expectations, and the quality dis closure should be complete, consistent, timely, accurate, reliable and comparable. An expectation gap may arise because o f deficient performance, deficient standards and users' unreasonable expectations. Content analysis is conducted to analyse the expectation gap. Overall, the findings are that GHG emissions disclosures do not meet the expectations of users. In relation to performance, man y cities have excluded several GHGs from emissions inventory, used multiple protocols to calculate emissions inventory, reported old emissions data and applied emissions reduction tar gets to a limited number of emissions sources. In relation to standards, the CDP guidelines are flexible regarding what is reported by cities, what protocols cities are using and there i s no checking for accuracy and completeness. In relation to expectations, that there are several geophysical and technical factors making it impossible to have comparable information at the city level. These findings suggest there is scope for the CDP to improve and standardise its disclosure system, working in collaboration with cities and the C40 alliance. Cities should also focus on improving their performance by providing accurate, complete, reliable and timely information. Even with standardisation, however, city-level emissions comparison is no t possible. A useful disclosure is therefore likely to be the actions cities take to reduce emissions and the impact of those actions. By disclosing actions and their impact, cities can learn best practices from each other and improve emissions reductions. The central recommendation is therefore that CDP should be asking for more information at a project or activity level from cities.A key finding from Paper 1 is that the differences between cities mean that the disclosure of emissions reduction actions will be much more useful for public accountability and peer learning than disclosure of raw emission levels. Paper 2 (Mia, et al., 2018a) investigates this issue further by exploring the quality of city emissions reduction actions and targets disclosures via traditional channels as well as social media. Three research questions are addressed. First, what communication channels are used by world megacities to dis close their emissions reduction targets and actions? Second, are these targets and ac tions communicated consistently across different channels? Third, what is the quality of the actions disclosed in different channels? Accountability theory is used to explain disclosures on the basis that cities are accountable to different groups of stakeholders to disclose information about their emissions reduction targets and actions which they can do by different channels. Media richness theory is used to explain why alternative disclosure might be useful e specially for discharging accountability. Ten megacities are selected and their emissions reduction targets and actions are examined across different channels to identify the common disclosure channels. Document analysis is conducted to assess the consistency of disclosed targets and actions related information, and a scoring system is developed to evaluate the quality of the disclosed actions related information across different channels. The findings are that cities have taken various actions including energy efficiency programs to reduce emission s and used multiple disclosure channels including social media. Disclosed information related to emissions reduction targets is consistent, but information about emissions reduction action s is mostly inconsistent, inadequate and mainly narrative in form. Concerning social medi a, this study finds that despite having a large number of online followers, several cities do no t use social media to disclose emissions reduction actions. The implication of these findings is that city authorities should better exploit the power of social media to inform and educate about their actions to reduce emissions, because it offers engagements and dialogue with the citizens to improve transparency and accountability. Standard setters and sustainability auditors also need to consider multiple disclosure channels to ensure information is consistent across the different channels to improve the credibility of the disclosed information.Paper 3 (Mia, et al., 2018b) extends the works of Papers 1 and 2 by focusing on one emissions reduction action in detail: energy efficiency. This particular action was selected as part of the findings of Paper 2 were that many cities including Australian cities are taking energy efficiency measures to reduce GHG emissions. Energy efficiency measure can reduce more than one-third of GHG emissions in the most cost-effective way (Magill, 2014; IEA, 2014). A comprehensive disclosure of the energy efficiency measures can help investors, community and other stakeholders to evaluate the cost and assess the social, economic and environmental benefits associated with energy efficiency measures. Paper 3 therefore explores the calculative practices and public disclosure about energy efficiency measure s within Australia's largest cities. From an accountability perspective, it is argued that Australian cities need to provide information about their energy efficiency activities in detail to the community and other stakeholders either to legitimise their existence or to fulfill the stakeholders' need or to help in their decision-making process. In the absence of any disclosure guideline for actions, a disclosure index is developed to assess the energy efficiency measures and applied to the eight major Australian cities in the sample, for the years 2017 and 2 018 (until 15 October). Findings of this study suggest that while many Australian cities are undertaking energy efficiency activities, they fall short in providing information related to costs and impacts of their energy efficiency measures and actions. Such limited disclosure makes it difficult for stakeholders to assess cities' energy efficiency measures and activities. These findings imply that a standardised and mandatory reporting requirements in relation t o energy efficiency (and other emission reduction projects) would facilitate enhanced transparency and accountability.Collectively, this thesis contributes to an understanding of the accounting and public disclosure practices of GHG emissions and reduction activities at the regional level. Beyond GHG emission disclosure research, in relation to other sustainability issues such as biodiversity or water, this thesis highlights that even with standardisation disclosures may not be comparable at the city or local government level because of different regional characteristics. Important disclosures for cities and similar type of agencies are the act ions taken to reduce sustainability viii harms and the impact of those actions. By making such disclosures, cities and local government can learn best practices from each other, with the aim of achieving sustainability. Therefore, policymakers and standard setters (e.g., CDP, GRI) should be as king for more information at a project or activity level. Also, standardised guidelines and verification processes are needed for actions related information to help organisations to provide complete, accurate, consistent, reliable and more comprehensive information. In addition, auditors should review the digital platform while conducting assurance service for traditional annual reports and SEA reports so that information is consistent across different channels. Standards setters also should contemplate the potentiality and popularity of digital platform s as these new form of communication channels offer both immediacy and interactivity which may be highly relevant in designing, developing and deploying SEA reporting frameworks.1 online resource (xii, 229 pages
Communication difficulty and effort in conversation
No full textTheoretical thesis.Bibliography: pages 143-159.Among the most important goals of hearing rehabilitation is the restoration of a hearing-impaired individual's capacity to take part in hearing tasks that form part of everyday life. This dissertation investigates the use of a new methodology for assessing the impacts of both hearing impairment and environmental noise on communication during spoken conversation. In order to balance the needs of ecological validity and experimental control, a tool for eliciting naturalistic conversations was designed and employed in conjunction with a set of highly realistic noise recordings. The development of this new approach to assessing communication effort is motivated by (i) the increasing recognition within hearing science that existing clinical and research measures may not adequately predict real-world hearing disability and hearing device benefit; and (ii) the increasing focus within the field on ecological validity and realistic assessment methods. A number of factors not typically considered in terms of ecological validity were considered. These include an emphasis on the importance of realistic tasks and the inherently interactive nature of spoken conversation. The extent to which tests of speech perception and hearing aid benefit approximate the real-life tasks people wish to accomplish and the cognitive demands of these tasks determines how accurately audiologists may measure hearing disability and predict real-world hearing device benefit and satisfaction. A study of conversation between young normal-hearing adults investigated the feasibility of using acoustic-phonetic measures of vocal effort to quantify the impacts of noise on communication difficulty. This was followed by an analysis of the naturalness of verbal behaviour elicited using these methods. Further, the subjective experience of participants during conversation was considered, including levels of engagement, and judgment of the difficulty and relevance of experimental conditions to real-life communication. A reduced set of acoustic-phonetic measures was employed to investigate the effects of age-related hearing impairment on conversational interaction and the novel concept of communication effort was proposed. Finally, the effect of hearing aid amplification on communication effort was investigated in terms of functional hearing disability and third-party disability experienced by communication partners. The findings of this study have implications for the definition of ecological validity within hearing science and provide a framework for measuring effort during conversation in order to better assess real-world hearing disability, third party disability of communication partners, and hearing device benefit for communication -- abstract.1. Introduction -- 2. Measuring communication difficulty through effortful speech production during conversation -- 3. Eliciting naturalistic conversations: a method for assessing communication ability, subjective experience and the impacts of noise and hearing impairment -- 4. Hearing impairment increases communication effort during conversations in noise -- 5. Hearing aid amplification reduces communication effort of people with hearing impairment and their conversation partners -- 6. Discussion, implications, and future directions.1 online resource (xxvi, 159 pages
Palestinian refugees’ views on the Right of return
No full textTheoretical thesis.Bibliography: pages 256-309.1. Introduction -- 2. Literature review -- 3. Research methodology -- 4. Socio-demographic characteristics of Palestinian refugees in the Occupied Territories and camps conditions -- 5. Right of return (Haq Al-Awda) -- 6. Compensation -- 7. The refugee camps and attitudes towards resettlement -- 8. The role of the United Nations relief and works agency for the Palestinian refugees -- 9. Conclusion and future directions -- 10. Reference list -- 11. Appendices.Right of return has been a fundamental claim by Palestinian people since 1948. The ‘right’ refers to the political position or principle that all generations of Palestinian refugees have the right to return to the property they or their forebears left behind during the 1948 Palestinian Exodus, and following the 1967 Six-Day War. Our current understanding of the Palestinian refugee perspective of the right of return however is limited due to the potential impact of changing socio-political conditions on Palestinian refugees’ views and the relative marginalisation of the refugee voice in the peace process.The aim of this study is to examine and update Palestinian refugees’ views of the right of return. Towards this aim, five aspects are given focus: the negotiation for the right of return; the centrality of the right of return claim to Palestinian refugees; compensation and the right of return claim; resettlement and the right of return claim; and the phase out of the United Nations Relief and Works Agency. This aim emerged directly in response to the marginalised Palestinian refugee voice and an interest in exploring the extent to which the right of return claim remains ‘sacred’ to them.Gaining access to interview refugees is challenging given the security situation and University ethics protocols. The study therefore adopted a quantitative research design including convenience sampling methods to recruit participants. The sample comprised 1200 participants from five refugee camps located in the West Bank. A self-administered survey instrument was used to facilitate the safe and confidential collection of data on the respondents’ views of a broad range of issues and social conditions. The survey was also designed to collect refugees’ demographic data (e.g., gender, age, education level attained etc.). This facilitated another key objective of this study; namely, to identify any significant relationships between demographic variables and refugee perceptions of the right of return claim or to its other key elements. Also considered in this thesis as a point of interest is the perceived ‘sacredness’ of the right of return to the Palestinian people as expressed in the literature.This study finds that most surveyed Palestinian refugees participants living in the West Bank are committed to the right of return as recommended in Resolution 194, and that most would not accept a peace agreement that did not include the provision for the right of return to their homeland. In addition, most surveyed Palestinian refugees reported that they would not accept compensation as a substitute for the right of return and just over half of all participants favoured resettlement in their original homes. Lastly, the surveyed Palestinian refugees overwhelmingly indicated that they wanted to remain aligned to the United Nations Relief and Works Agency until the right of return is actioned.The main conclusion drawn from these findings is that, even after 70 years of displacement, the right of return remains an active but changing political construct among Palestinians in refugee camps located in the West Bank. Furthermore, based on the findings indicating most respondents remain committed to Resolution 194 and demonstrate a general reluctance towards accepting compensation or resettlement as a substitute, this study concludes that the right of return claim maintains a sense of sacredness among the surveyed refugees. As such, future negotiations must consider the generational narratives and ensure that the right of return claim, resettlement, and compensation particularly are not treated as mutually exclusive in the delivery of a just solution to the displacement of Palestinian refugees.Mode of access: World wide web1 online resource (351 pages) diagrams, graph
Structures in two dimensional category theory and applications to polynomial functors
No full textTheoretical thesis."Centre of Australian Category Theory Department" -- title page.Bibliography: pages 249-253.1. Introduction -- 2. Yoneda structures and KZ doctrines -- 3. Distributive laws via admissibility -- 4. Generic bicategories -- 5. Universal properties of bicategories of polynomials -- 6. An elementary view of familial pseudofunctors -- 7. Conclusion and future directions -- References.This thesis seeks to further develop two-dimensional category theory, with a focus on Yoneda structures, (lax-idempotent) pseudomonads, pseudo-distributive laws, and familial representability, in order to gain new insights and tools in the study of polynomial functors.The first contribution of this thesis concerns Yoneda structures, which give a formalization of the presheaf construction. Our main result shows that any fully faithful lax-idempotent pseudomonad almost gives rise to a Yoneda structure, with all of the axioms holding except for one condition.The second contribution of this thesis concerns pseudo-distributive laws of a pseudomonad and a lax-idempotent pseudomonad. We show that such distributive laws have a simple algebraic description which only requires three out the usual eight coherence conditions, and another simple description in terms of the data of the near-Yoneda structurere covered from the lax-idempotent pseudomonad.Our third contribution is to introduce a class of bicategories, which we term generic bicategories. These are the bicategories for which horizontal composition admits generic factorisations, and have the interesting property that oplax functors out of them have a reduced description, similar to the axioms of a comonad.The fourth contribution of this thesis is to establish the universal properties of the bicategory of polynomials, with general and cartesian 2-cells, using the properties of generic bicategories to avoid the majority of the coherence conditions. In addition, we give a new proof of the universal properties of the bicategory of spans and establish the universal properties of the bicategory of spans with invertible 2-cells.The fifth contribution of this thesis is to give an appropriate notion of familial representability for pseudofunctors L : A - B of bicategories, and to describe an equivalence with an analogue of generic factorisations. This improves on work of Weber, who did not provide such an equivalence, and required A to have a terminal object.Mode of access: World wide web1 online resource (xii, 253 pages
Genre and impact captioning in a Korean real-variety show: a systemic functional multimodal discourse analysis
No full textTheoretical thesis.Bibliography: pages 134-148.Chapter 1. Introduction -- Chapter 2. Literature review -- Chapter 3. Methodology and data collection -- Chapter 4. A genre approach : analysis on 3 meals a day season 6 -- Chapter 5. Impact captioning -- Chapter 6. Conclusion -- References -- Appendices.This thesis aims to identify how meaning is made in one popular Korean TV Real Variety Show (RVS) from a systemic functional multimodal discourse analysis (SF-MDA) perspective, which has not been widely used in Korean academia. The focus of this paper is on analysis of structure, and impact captions of 3 Meals a Day (3 MaD) season 6. Impact captions (ICs) are new forms of writing on TV entertainment shows. Studies of ICs have continued for a few decades, mostly in Korea and Japan. Yet, this research area is still new to the West both in the TV industry and academia. This paper attempts to contribute to existing IC studies by adding a relatively new perspective, or an SF-MDA perspective, to look into Korean RVSs and ICs. The scope of data in this paper is limited to one season (season 6) out of six for the structure analysis, and four episodes (episode 1, 3, 4, and 5) for the IC analysis. Methodologically, genre theory and systemic functional theory are applied in the analysis of the data. The results of this analysis are summarised in two parts as follows. First, on the basis of the analysis of the structure, there is a functional structure of 3 MaD season 6 which uncovers the relations between each component or stage, and presents how meaning is made through these syntagmatic relations. Second, the analysis of the ICs examines how meaning is made by a selection of systemic choices. This paper identifies system networks of ICs at the level of expression, opening a way to construct system networks at the level of content.Mode of access: World wide web1 online resource (x, 203 pages) colour illustration
The Mel-frequency cepstrum coefficient for music emotion recognition in machine learning
No full textEmpirical thesis.Bibliography: pages 45-49.1. Introduction -- 2. Music features -- 3. Emotion models -- 4. Mel-frequency cepstrum coefficient -- 5. Temporal and spectral properties -- 6. Existing work in music emotion recognition -- 7. The experiment -- 8. Results and discussion -- 9. References.The Mel-frequency Cepstrum Coefficient (MFCC), a technique designed initially for speech analysis, has in recent years become very popular in music emotion recognition projects. MFCC uses the Mel scaling method to simulate human auditory properties, logarithmic noise reduction techniques, and the Discrete Cosine Transformation (DCT) to generalise all salient features, without losing critical information. These techniques, while applicable to speech analysis, may not always be suitable for music analysis. We suggest, in Music Emotion Recognition (MER) analysis, spectral and temporal (which have a deep historical foundation) should be the more relevant features to use.We propose extracting three feature types, MFCC, Spectral, and Temporal, from the clips of songs in the '1000 songs' dataset to train a simple Artificial neural network (ANN). The trained ANN model will subsequently be able to predict the emotion value of songs. The prediction error is calculated based on the predicted value and actual annotated value. The feature that produces the lowest prediction error is judged as the most suitable feature for MER. Our results show that spectral features produced the lowest error, whereas MFCC produced the highest prediction error; this suggests that MFCC may not be a suitable feature for MER.Mode of access: World wide web1 online resource (x, 47 pages) diagrams, graphs, table
A reduced order approach for probabilistic inversions of 3D magnetotelluric data
No full textEmpirical thesis."Australian Research Council Centre of Excellence for Core to Crust Fluid Systems (CCFS) and GEMOC" -- title page.Bibliography: pages 183-199.1. Introduction -- 2. Fundamentals of the magnetotelluric method -- 3. Solution of the magnetotelluric forward problem -- 4. Probabilistic inverse problems -- 5. A combined reduced basis-Markov chain Monte Carlo method for large-scale MT inversion problems -- 6. Numerical examples -- 7. Conclusions and future work -- Appendix -- References.Multi-observable probabilistic inversion (e.g. Afonso et al., 2013a,b, 2016b) is a recent framework specifically designed to provide insights into the physicochemical structure of the lithosphere and its complex interactions with the sublithospheric upper mantle. Of particular relevance is the inclusion of 3D magnetotelluric (MT) data, as it provides complementary information not only on the thermal structure but also on water content and fluid pathways; this is critical for understanding and imaging the complex fluid-rock interactions responsible for mineralization events and water-assisted tectonism. However, in order to isolate the effect of fluids from other potential compositional and thermal \background effects, MT data needs to be informed by other data sets such as seismic and gravity data.In order to include MT data into multi-observable probabilistic inversions, we first need to solve the problem of computational effciency when solving the MT equations in 3D. For this, we have combined probabilistic inversion methods, parallel MT solvers (Zyserman & Santos, 2000) and advanced reduced order modelling techniques to obtain fast, yet accurate, solutions to both the MT inversion and the full 3D joint inversion of MT and surface wave data. Such a probabilistic formalism offers a natural framework to assess non-uniqueness and uncertainties affecting the inversion, which are otherwise hard to quantify using traditional inversion methods.The outcomes of this thesis demonstrate the capabilities of the conceptual and numerical framework for 3D multi-observable probabilistic inversions and open up new exciting opportunities for integrated geophysical imaging of the Earth's interior.Mode of access: World wide web1 online resource (xx, 199 pages) diagrams, graphs, table