6297 research outputs found
Sort by
INVESTIGATION OF ORTHOGONAL POLYNOMIAL KERNELS AS SIMILARITY FUNCTIONS FOR PATTERN CLASSIFICATION BY SUPPORT VECTOR MACHINES
Get PDFA kernel function is an important component in the support vector machine (SVM) kernel-based classifier. This is due to the elegant mathematical characteristics of a kernel, which amount to the mapping of non-linearly separable classes to an implicit higher-dimensional feature space where they can become linearly separable, and hence easier to classify. Such characteristics are those prescribed by the underpinning positive semi-definite (PSD) property. The properties of this feature space can, however, be difficult to interpret, to customize or select an appropriate kernel for the classification task at hand. Moreover, the high-dimensionality of the feature space does not usually provide apparent and intuitive information about the natural representations of the data in the input space, as the construction of this feature space is only implicit. On the other hand, SVM kernels have also been regarded as similarity functions in many contexts to measure the resemblance between two patterns, which can be from the same or different classes. However, despite the elegant theory of PSD kernels, and its remarkable implications on the performance of many learning algorithms, limited research efforts seem to have studied kernels from this similarity perspective. Given that patterns from the same class share more similar characteristics than those belonging to different classes, this similarity perspective can therefore provide more tangible means to craft or select appropriate kernels than the properties of the implicit high-dimensional feature spaces that one might not even be able to calculate.
This thesis therefore aims to: (i) investigate the similarity-based properties, which can be exploited to characterise kernels (with focus on the so-called “orthogonal polynomial kernels”) when used as similarity functions, and (ii) assess the influence of these properties on the performance of the SVM classifier. An appropriate similarity-based model is therefore defined in the thesis based on how the shape of an SVM kernel should ideally look like when used to measure the similarity between its two inputs. The model proposes that the similarity curve should be maximized when the two kernel inputs are identical, and it should decay monotonically as they differ more and more from each other. Motivated by the pictorial characteristics of the Chebyshev kernels reported in the literature, the thesis adopts this kernel-shape perspective to also study some other orthogonal polynomial kernels (such as the Legendre kernels and Hermite kernels), to underpin the assessment of the proposed ideal shape of the similarity curve for kernel-based pattern classification by SVMs.
The analysis of these polynomial kernels revealed that they are naturally constructed from smaller kernel building blocks, which are combined by summation and multiplication operations. A novel similarity fusion framework is therefore developed in this thesis to investigate the effect of these fusion operations on the shape characteristics of the kernels and on their classification performance. This framework is developed in three stages, where Stage 1 kernels are those building blocks constructed from only the polynomial order n (the highest order under consideration), whereas Stage 2 kernels combine all the Stage 1 kernel blocks (from order 0 to n) using a summation fusion operation. The Stage 3 kernels finally combine Stage 2 kernels with another kernel via a multiplication fusion operation. The analysis of the shape characteristics of these three-stage polynomial kernels revealed that their inherent fusion operations are synergistic in nature, as they bring their shapes closer to the ideal similarity function model, and hence enable the calculation of more accurate similarity measures, and accordingly score better classification performance. Experimental results showed that these summative and multiplicative fusion operations improved the classification accuracy by average factors of 17.35% and 19.16%, respectively, depending on the dataset and the polynomial function employed.
On the other hand, the shapes of the Stage 2 polynomial kernels have also been shown to oscillate after a certain threshold within the standard normalized input space of [-1,1]. A simple adaptive data normalization approach is therefore proposed to confine the data to the threshold window where these kernels exhibit the sought after ideal shape characteristics, hence eliminate the possibility of any data point to be located outside the range where these oscillations are observed. The implementation of the adaptive data normalization approach accordingly leads to a more accurate calculation of similarity measures and improves the classification performance. When compared to the standard normalized input space, experimental results (performed on the Stage 2 kernels) demonstrate the effectiveness of the proposed adaptive data normalization approach, with an average accuracy improvement factor of 11.772%, depending on the dataset and the polynomial function utilized.
Finally, a new perspective is also introduced whereby the utilization of orthogonal polynomials is perceived as a way of transforming the input space to another vector space, of the same dimensionality as the input space, prior to the kernel calculation step. Based on this perspective, a novel processing approach, based on vector concatenation, is proposed which, unlike the previous approaches, ensures that the quantities processed by each polynomial order are always formulated in vector form. This way, the attributes embedded in the structure of the original vectors are maintained intact. The proposed concatenated processing approach can also be used with any polynomial function, regardless of the parity combination of its monomials, whether they are only odd, only even, or a combination of both. Moreover, the Gaussian kernel is also proposed to be evaluated on vectors processed by the polynomial kernels (instead of the linear kernel used in the previous approaches), due to the more accurate similarity shape characteristics of the Gaussian kernel, as well as its renowned ability to implicitly map the input space to a feature space of higher dimensionality. Experimental results demonstrate the superiority of the concatenated approach for all the three polynomial-kernel stages of the developed similarity fusion framework and for all the polynomial functions under investigation. When the Gaussian kernel is evaluated on the vectors processed using the concatenated approach, the observed results show a statistically significant improvement in the average classification accuracy of 22.269%, compared to when the linear kernel is evaluated on the vectors processed using the previously proposed approaches
Coping Strategies, Psychological Impact, and Support Preferences of Men With Rheumatoid Arthritis: A Multicenter Survey
Get PDFObjective
To investigate the existence and distribution of 2 typologies (termed “factors”) of men with rheumatoid arthritis (RA) identified through our previous Q‐methodology study (n = 30) in a larger sample of men with RA, and whether differences in psychosocial impact or support preferences exist between the 2 factors, and between men and women with RA.
Methods
A postal survey was sent to 620 men with RA from 6 rheumatology units across England, and the support preferences section of the survey was given to 232 women with RA.
Results
A total of 295 male patients (47.6%) and 103 female patients (44.4%) responded; 15 male participants had missing data, and thus 280 were included in the analysis. Of these, 61 (22%) were assigned to factor A (“accept and adapt”), 120 (35%) were assigned to factor B (“struggling to match up”), and 99 (35%) were unassigned. The two factors differed significantly, with factor B reporting more severe disease, less effective coping strategies, and poorer psychological status. For support, men favored a question and answer session with a consultant (54%) or specialist nurse (50%), a website for information (69%), a talk by researchers (54%), or a symptom management session (54%). Overall, women reported more interest in support sessions than men, with ≥50% of women reporting interest in nearly every option provided.
Conclusion
Some men accept and adapt to their RA, but others (43%) report severe disease, less effective coping, and poor psychological status. Men's preferences for support are practical, with a focus on expanding their knowledge
How are you feeling?’ Recognising Postnatal Depression
No full textPostnatal Depression (PND) in Black, Asian and Minority Ethnic (BAME) women in the UK is higher due to additional risk factors: such as new bride; new culture; unfamiliar antenatal care; poor housing; unemployment lack of interpreters and poor communication.
BAME women are also less likely to be identified as having PND and so remain undetected and untreated.
The ‘How are you feeling?’ picture booklets were designed to provide culturally appropriate images and language to facilitate communication and diagnosis of PND in women whose first language is
Urdu, Bengali, Somali, Chinese, Arabic and English
Conventional and Alternative Power Generation
No full textThis book goes beyond the traditional methods of power generation. It introduces the many recent innovations on the production of electricity and the way they play a major role in combating global warming and improving the efficiency of generation. It contains a strong analytical approach to underpin the theory of power plants—for those using conventional fuels, as well as those using renewable fuels—and looks at the problems from a unique environmental engineering perspective. The book also includes numerous worked examples and case studies to demonstrate the working principles of these systems.
Conventional and Alternative Power Generation: Thermodynamics, Mitigation and Sustainability is divided into 8 chapters that comprehensively cover: thermodynamic systems; vapor power cycles, gas power cycles, combustion; control of particulates; carbon capture and storage; air pollution dispersal; and renewable energy and power plants.
Features an abundance of worked examples and tutorials
Examines the problems of generating power from an environmental engineering perspective
Includes all of the latest information, technology, theories, and principles on power generation
Conventional and Alternative Power Generation: Thermodynamics, Mitigation and Sustainability is an ideal text for courses on mechanical, chemical, and electrical engineering
Peak-to-average power ratio reduction for DCO-OFDM underwater optical wireless communication system based on an interleaving technique
Get PDFIn underwater direct current-biased optical orthogonal frequency-division multiplexing (DCO-OFDM) system, high peak-to-average power ratio (PAPR) brings in-band distortion and out-of-band power. It also decreases the signal-to-quantization noise ratio of the analog-to-digital converter and the digital-to-analog converter. A time–frequency-domain interleaved subbanding scheme is proposed to reduce the PAPR of underwater DCO-OFDM system with low computation complexity and bit error rate (BER). The system BER is evaluated by the distances of the underwater optical wireless communication (UOWC), as well as by the signal attenuation of the UOWC channel. A least-square channel estimation method is adopted for adaptive power amplification at the receiver side, to further decrease the system BER
Transforming the stakeholders’ Big Data for intellectual capital management
Get PDFPurpose
This paper aims to define a conceptual framework for transforming Big Data into organizational value by focussing on the perspectives of service science and activity theory. In coherence with the agenda on evolutionary research on intellectual capital (IC), the study also provides momentum for researchers and scholars to explore emerging trends and implications of Big Data for IC management.
Design/methodology/approach
The paper adopts a qualitative and integrated research method based on a constructive review of existing literature related to IC management, Big Data, service science and activity theory to identify features and processes of a conceptual framework emerging at the intersection of previously identified research topics.
Findings
The proposed framework harnesses the power of Big Data, collectively created by the engagement of multiple stakeholders based on the concepts of service ecosystems, by using activity theory. The transformation of Big Data for IC management addresses the process of value creation based on a set of critical dimensions useful to identify goals, main actors and stakeholders, processes and motivations.
Research limitations/implications
The paper indicates how organizational values can be created from Big Data through the co-creation of value in service ecosystems. Activity theory is used as theoretical lens to support IC ecosystem development. This research is exploratory; the framework offers opportunities for refinement and can be used to spearhead directions for future research.
Practical implications
The paper proposes a framework for transforming Big Data into organizational values for IC management in the context of entrepreneurial universities as pivotal contexts of observation that can be replicated in different fields. The framework provides guidelines that can be used to help organizations intending to embark on the emerging paradigm of Big Data for IC management for their competitive advantages.
Originality/value
The paper’s originality is in bringing together research from Big Data, value co-creation from service ecosystems and activity theory to address the complex issues involved in IC management. A further element of originality offered involves integrating such multidisciplinary perspectives as a lens for shaping the complex process of value creation from Big Data in relationship to IC management. The concept of how IC ecosystems can be designed is also introduce
Analysis of Brazilian fashion sectorial brand identity
Get PDFThe present study proposes to identify similar characteristics and managerial actions of the sector brand identity elements among the associations that compound the Brasil Fashion System brand.
An exploratory qualitative research was developed through in-depth interviews conducted with associations of the Brazilian fashion sector.
The results indicate that there are characteristics of the elements of brand identity that are similar between the associations that compound the Brasil Fashion System brand. However, there are also several distinct characteristics among them, which makes it difficult, in large part, to consolidate the brand identity of the Brazilian fashion industry abroad.
Moreover, it was indicated that for sectorial brand cases with a great divergence among brand partners, the creation of sub-sectorial brand specific for each partner could bring better results, since in this way brands could be created with more suitable attributes for each partner, which will better suit their target audiences.
A practical contribution is also obtained since the study can help in the elaboration of improvements for the sectorial brands that represents a large partners group An empirical evidence of how identify common attributes between sectoral brand partners was presented in order to have a consolidated brand image in external market
“This is proof”? Forensic evidence and ambiguous material culture at Treblinka extermination camp
Get PDFIn recent years, a forensic archaeological project at Treblinka extermination camp has uncovered significance evidence relating to the mass murder that took place there. A number of questions emerged regarding the provenance and origins of objects discovered as part of this work, and why they had remained undiscovered for over seventy years. These discoveries led to an opportunity to confirm and challenge the history of the extermination camp, and demands (from the public) to view the objects. This paper will outline how archaeologists and artists came together to reflect on these issues, whilst simultaneously providing access to the new findings
Activating the unemployed through Sociedades Laborales in Spain
No full textSpanish Sociedades Laborales (SLs) – mostly small and micro enterprises – are a qualified form of conventional corporation, majority-owned by their permanent employees. Unemployed persons can capitalise their unemployment benefits as a lump sum to start a new SL or to recapitalise an existing SL by joining it. This makes SLs similar to start-up subsidies for the unemployed, an established instrument of active labour market policy across the EU. The new 2015 Law on Worker-Owned and Participatory Companies substantially modernised the concept of SLs 30 years after its inception.
SLs provide an unemployed person who joins or sets up an SL not only with access to capital but with business and entrepreneurial mentoring and practical expertise. These enterprises also play an important role in job creation and expanding secondary employment. They are based on employee ownership whose demonstrated benefits complement the policy aims of ALMPs. While there are no obstacles to transfering the Sociedad Laboral to other Member States, the model has important benefits that make it particularly suitable for combination with existing national start-up incentives for the unemployed.
This report investigates the potential of SLs as an instrument of ALMP for returning the unemployed to the labor market and also the transferability of the scheme to other EU Member States
In-house, University-based work experience versus off-campus, work-experience
No full textPurpose: To investigate students’ perceptions of the value, impact, benefits and disadvantages of in-house, University-based work experience versus off-campus, work-experience.
Design/methodology/approach: Three focus groups, one consisting of students who had undertaken work experience off-campus at an employers’ workplace (n=6), one consisting of students who had undertaken work experience in-house with a University-based employer (n=6), and a third mixed group (n=6, consisting of students who had undertaken both types), were formed. Focus group data were supplemented by interviews (n=3). Data were transcribed and analysed thematically.
Findings: Based on student perceptions, both types of work experience were thought to: enhance future employment; provide career insight; enable skill/experience acquisition and application; and be useful for building relationships. Work experience that occurred in-house was, in addition, perceived to: be cost effective; enable students to be more closely supervised and supported; be good for relationship building between and within students/staff; be beneficial for increasing student attainment; and enable students to see the link between theory and practice more clearly. In-house work experience was, however, deemed to be restricted in terms of variety, and links with and perceptions of external stakeholders.
Research limitations/implications: The study is limited in that it is based on the perceptions of students undertaking unique types of integrated work experience within one faculty at one university.
Practical implications: When deciding on whether in-house or off-campus work experiences are offered, consideration should be given to level of support, supervision, observation, and travel and time costs.
Originality/value:Original views of students regarding in-house work experience have been gathered, which can be used to inform in-course workplace practices
Data sources:
Access Repository Dashboard
Do you manage STORE - Staffordshire Online Repository? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!