1,721,185 research outputs found
Use of Data Mining in System Development Life Cycle
No full textThis volume provides a snapshot of the current state of the art in data mining, presenting it both in terms of technical developments and industrial applications. The collection of chapters is based on works presented at the Australasian Data Mining conferences and industrial forums. Authors include some of Australia's leading researchers and practitioners in data mining. The volume also contains chapters by regional and international authors. The original papers were initially reviewed for the workshops, conferences and forums.\ud
\ud
The 25 articles in this state-of-the-art survey were carefully reviewed and selected from numerous contributions during at least two rounds of reviewing and improvement for inclusion in the book. They provide an interesting and broad update on current research and development in data mining. The book is divided into two parts. It starts with state-of-the-art research papers organized in topical sections on methodological advances, data linkage, text mining, and temporal and sequence mining. The second part comprises papers on state-of-the-art industrial applications from the fields of health, finance and retail
Development and leadership in computer-mediated collaborative groups
Get PDFComputer-mediated collaboration is an important feature of modern organisational and educational settings. Despite its ever increasing popularity, it is still commonly compared unfavourably with face-to-face collaboration because non-verbal and paralinguistic cues are minimal. Although research on face-to-face group collaboration is well documented, less is known about computer-mediated collaboration.
The initial focus of this thesis was an in-depth analysis of a case study of a computer-mediated collaborative group. The case study was a large international group of volunteer researchers who collaborated on a two-year research project using asynchronous communication (email). This case study was a window on collaborative dialogue in the early 1990s (1992-94) at a time when information and communication technologies were at an early stage of development.
After identifying the issues emerging from this early case study, another case study using technologies and virtual environments developed over the past decade, was designed to further understand how groups work together on a collaborative activity. The second case study was a small group of students enrolled in a unit of study at Murdoch University who collaborated on a series of nine online workshops using synchronous communication (chat room). This case study was a window on collaborative dialogue in the year 2000 when information and communication technologies had developed at a rate which few people envisioned in the early 90s.
The primary aim of the research described in this thesis was to gain a better understanding of how computer-mediated collaborative communities develop and grow. In particular, the thesis addresses questions related to the developmental and leadership characteristics of collaborative groups.
Internet research requires a set of assumptions relating to ontology, epistemology, human nature and methodological approach that differs from traditional research assumptions. A research framework for Internet research - Complementary Explorative Data Analysis (CEDA) - was therefore developed and applied to the two case studies.
The results of the two case studies using the CEDA methodology indicate that computer-mediated collaborative groups are highly adaptive to the aim of the collaborative task to be completed, and the medium in which they collaborate. In the organisational setting, it has been found that virtual teams can devise and complete a collaborative task entirely online. It may be an advantage, but it is certainly not mandatory to have preliminary face-to-face discussions. What is more important is to ensure that time is allowed for an initial period of structuration which involves social interaction to develop a social presence and eventually cohesiveness. In the educational setting, a collaborative community increases pedagogical effectiveness. Providing collaborative projects and interdependent tasks promotes constructivist learning and a strong foundation for understanding how to collaborate in the global workplace. Again, this research has demonstrated that students can collaborate entirely online, although more pedagogical scaffolding may be required than in the organisational setting. The importance of initial social interaction to foster a sense of presence and community in a mediated environment has also been highlighted.
This research also provided greater understanding of emergent leadership in computer-mediated collaborative groups. It was found that sheer volume of words does not make a leader but frequent messages with topic-related content does contribute to leadership qualities.
The results described in this thesis have practical implications for managers of virtual teams and educators in e-learning
A Framework for Customisable Sport Video Management and Retrieval
No full textSeveral domain specific approaches for video management have shown the benefits of integrating low- and high- level video contents in supporting more robust retrievals. However, there are not many work has shown how to integrate them in order to support different types of video. In this paper, we firstly propose a framework for customisable video management system which allows the system to detect the type of video to be indexed, so that appropriate tools can be used to extract the key segments. It is also customisable because the system manages user preferences and usage history to make the system supports specific requirements. Secondly, we will show how the extracted key segments can be summarised using standard descriptions of MPEG-7 in a hierarchical scheme which is potentially easy to share between users. Thirdly, we have developed and tested some queries which show that XQuery provides a powerful language for our video management’s retrieval
Computational intelligent data analysis for sustainable development : an introduction and overview
No full textThe concept of sustainability received worldwide recognition as a result of a report that was published in 1987 by the World Commission on Environment and Development (known as the Brundtland Commission), titled “Our Common Future”. The commission developed today’s generally accepted definition of sustainability, stating that sustainable development is development that meets the needs of the present without compromising the ability of future generations to meet their own needs. The three main pillars of sustainable development include economic growth, environmental protection, and socio-political sustainability. While many people agree that each of these three ideas contributes to the overall idea of sustainability, it is difficult to find evidence of equal levels of initiatives for the three pillars in governmental policies worldwide
Visual communication and trust in the health domain
No full textVisual communication and user trust are always challenges in the health domain where conservativeness, precision, and domain knowledge can outweigh the validity of the outcomes in the analytical methods and processes. It is crucial to provide better awareness and understanding to domain experts, model developers, and even patients who might be conscious of their condition and how a treatment or a diagnosis is decided for them. This chapter contributes a discussion on trust and its issues in health data-driven science and how trust should be associated with analytical and computational processes, which are enhanced by visualisation and interaction. We also provide brief guidance on the models and methods for improving interpretability and trust in the health domain
Style Recognition using Keyword Analysis
No full textThe primary aim of this research project is to develop a generic framework and methodologies that will enable the augmentation of expert knowledge with knowledge extracted from multimedia sources such as text and pictures, for the purpose of classification and analysis. For evaluation and testing purposes of this research study, a furniture design style domain is selected because it is a common belief that design style is an intangible concept that is difficult to analyze. In this paper, we present the results of the analysis of keywords in the text descriptions of design styles. A simple keyword-based matching technique is used for classification and domain specific dictionaries of keywords are used to reduce the dimensionality of feature space. A comparative evaluation was carried out for this classifier and SVM and decision tree based classifier C4.
Visualisation for explainable machine learning in biomedical data analysis
No full textThis chapter covers innovations in biomedical data mining and interpretations, especially using visualisations in interpretable machine learning for biomedical data analysis. Visualisations are important in presenting artificial intelligence models and validating the machine learning results. There are more new and complex machine learning methods that have been created to assist decision-making in recent years in the medical domain. Most of them are treated as “black boxes”, as the training and prediction processes are hidden in complicated mathematical theories. Visualisation is a way to reveal the process and help a human understand the cause of a decision. Knowing the “why” for the prediction results and “how” the model works can improve users’ trust in artificial intelligence results. The chapter introduces different visualisations used in interpreting supervised and unsupervised machine learning models for biomedical data. We also provide discussions and future work on using visualisations in interpreting data mining results in the medical domain
Feature-ranking methods for RNA sequencing data
No full textRibonucleic acid sequencing (RNA-Seq) is a technique that is used a lot to study and evaluate gene expression patterns and find genes that are expressed differently in different biological situations. Numerous computational algorithms for analysing RNA-seq data have been developed that categorise them per features in many pre-defined classifications. Feature-ranking techniques have emerged as a powerful tool for analysing RNA sequencing data, enabling the identification of the most relevant genes that are associated with specific phenotypes or biological processes. In this chapter, we give an overview of different ways to rank features and how they can be used to analyse data from RNA sequencing. We also compare how well different methods work using benchmark datasets and talk about the difficulties of combining multiple data sources and figuring out what the results mean. Last, we talk about possible future directions for the development and use of feature-ranking techniques. These include the use of deep learning techniques, the use of single-cell sequencing data, and the development of methods for figuring out how genes interact with each other. We evaluate selected features by optimising parameters and identifying a higher-performing classifier. The accuracy, recall, false-positive rate (FPR), and precision are used to analyse the comparison. The chapter aims to provide a comprehensive guide for researchers who want to use feature-ranking techniques to analyse RNA sequencing data and gain insights into the underlying biology
Assisting Human Cognition in Visual Data Mining
Get PDFAs discussed in Part 1 of the book in chapter Form-Semantics-Function. A Framework for Designing Visualisation Models for Visual Data Mining the development of consistent visualisation techniques requires systematic approach related to the tasks of the visual data mining process. Chapter Visual discovery of network patterns of interaction between attributes presents a methodology based on viewing visual data mining as a reflection-in-action process. This chapter follows the same perspective and focuses on the subjective bias that may appear in visual data mining. The work is motivated by the fact that visual, though very attractive, means also subjective, and non-experts are often left to utilise visualisation methods (as an understandable alternative to the highly complex statistical approaches) without the ability to understand their applicability and limitations. The chapter presents two strategies addressing the subjective bias: guided cognition and validated cognition, which result in two types of visual data mining techniques: interaction with visual data representations, mediated by statistical techniques, and validation of the hypotheses coming as an output of the visual analysis through another analytics method, respectively
- …
