The recent decade has witnessed great advances in microarray and genotyping technologies which allow genome-wide single nucleotide polymorphism (SNP) data to be captured on a single chip. As a consequence, genome-wide association studies require the development of algorithms capable of manipulating ultra-large-scale SNP datasets. Towards this goal, this thesis proposes two SNP selection methods; the first using Independent Component Analysis (ICA) and the second based on a modified version of Fast Orthogonal Search. The first proposed technique, based on ICA, is a filtering technique; it reduces the number of SNPs in a dataset, without the need for any class labels. The second proposed technique, orthogonal search based SNP selection, is a multivariate regression approach; it selects the most informative features in SNP data to accurately model the entire dataset. The proposed methods are evaluated by applying them to publicly available gene SNP datasets, and comparing the accuracies of each method in reconstructing the datasets. In addition, the selection results are compared with those of another SNP selection method based on Principal Component Analysis (PCA), which was also applied to the same datasets. The results demonstrate the ability of orthogonal search to capture a higher amount of information than ICA SNP selection approach, all while using a smaller number of SNPs. Furthermore, SNP reconstruction accuracies using the proposed ICA methodology demonstrated the ability to summarize a greater or equivalent amount of information in comparison with the amount of information captured by the PCA-based technique reported in the literature. The execution time of the second developed methodology, mFOS, has paved the way for its application to large-scale genome wide datasets. / Thesis (Master, Computing) -- Queen's University, 2010-12-15 18:03:00.208

This thesis investigates the application of principal component analysis to the Australian stock market using ASX200 index and its constituents from April 2000 to February 2014. The first ten principal components were retained to present the major risk sources in the stock market. We constructed portfolio based on each of the ten principal components and named these “principal portfolios

Abstract Beliefs about which behaviors and responsibilities should typical be assumed by women and men are central in shaping gender relations and gender equality in society. The belief that women should be responsible for domestic work, while men should provide economically for the family gives rise to an uneven opportunity structure, situating women in a disadvantaged position compared to men. In order to achieve gender equality traditional gender role attitudes need to liberalize. This thesis examines who supports non-traditional gender roles in Sweden. Data representative of the Swedish population between the ages of 18-79 were used to explore the relationship between social context and individual self-interest and gender role attitudes. The results showed that women are more likely to be positive towards non-traditional gender roles if they are situated in highly educated social contexts. Conversely, men were found to be more likely to be positive if situated in gender equal contexts. This indicates that men’s beliefs regarding what is appropriate for women might be countered by women in gender equal contexts, while women may find confirmation regarding their non-traditional gender role attitude in other equally liberal women.

Substance abuse is the overindulgence in and dependence on a drug or chemical leading to detrimental effects on the individual’s health and the welfare of those surrounding him or her. Logistic regression analysis is an important tool used in the analysis of the relationship between various explanatory variables and nominal response variables. The objective of this study is to use this statistical method to determine the factors which are considered to be significant contributors to the use or abuse of substances in school-aged children and also determine what measures can be implemented to minimize their effect. The logistic regression model was used to build models for the three main types of substances used in this study; Tobacco, Alcohol and Drugs and this facilitated the identification of the significant factors which seem to influence their use in children.

A novel imaging technology, digital breast tomosynthesis (DBT), is a technique that overcomes the tissue superposition limitation of conventional mammography by acquiring a limited number of X-ray projections from a narrow angular range, and combining these projections to reconstruct a pseudo-3D image. The emergence of DBT as a potential replacement or adjunct to mammographic screening mandates that solutions be found to two of its major limitations, namely X-ray scatter and mono-energetic reconstruction methods. A multi-faceted software-based approach to enhance the image quality of DBT imaging has the potential to increase the sensitivity and specificity of breast cancer detection and diagnosis. A scatter correction (SC) algorithm and a spectral reconstruction (SR) algorithm are both ready for implementation and clinical evaluation in a DBT system and exhibit the potential to improve image quality. A principal component analysis (PCA) based model of breast shape and a PCA model of X-ray scatter optimize the SC algorithm for the clinical realm. In addition, a comprehensive dosimetric characterization of a FDA approved DBT system has also been performed, and the feasibility of a new dual-spectrum, single-acquisition DBT imaging technique has also been evaluated.

Behavioural syndromes, defined as suites of correlated behaviours across different contexts, are used to characterize individual variability in behaviours. Males of the weakly electric fish species, Apteronotus leptorhynchus, produce electro-communication signals called chirps. Chirps are thought to be involved in agonistic signalling, as their relative incidence increases during agonistic conspecific interactions. However, high levels of individual variability in aggression obscure the role of chirps in mediating aggression. Here, I tested the presence of an aggression-boldness behavioural syndrome, and then considered the implications such a syndrome would have on chirping behaviours. Behavioural tests in anti-predation, object novelty, feeding, conspecific intrusion and novel environment exploration contexts revealed a syndrome involving only object novelty and feeding. We found no correlation between chirping behaviour and the assessed behaviours. Our results demonstrate that chirps represent a more complex communication system than previously suggested.

Research into traumatic brain injury (TBI) mechanisms is essential for the development of methods to prevent its occurrence. One of the most common ways to incur a TBI is from falls, especially for the young and very old. The purpose of this thesis was to investigate how the acceleration loading curves influenced the occurrence of different types of TBI, namely: epidural hematoma, subdural hematoma, subarachnoid hemorrhage, and contusion. This investigation was conducted in three parts. The first study conducted reconstructions of 20 TBI cases with varying outcomes using MADYMO, Hybrid III, and finite element methodologies. This study provided a dataset of threshold values for each of the TBI injuries measured in parameters of strain and stress. The results of this study indicated that using a combined reconstructive approach produces results which are in keeping with the literature for TBI. The second study examined how the characteristics of the loading curves which were produced from each reconstruction influenced the outcome using a principal components analysis. It was found that the duration of the event accounted for much of the variance in the results, followed with the acceleration components. Different curve characteristics also accounted for differing amounts of variance in each of the lesion types. Study 3 examined how the dynamic response of the impact influenced where in the brain a subdural hematoma (SDH) could occur. It was found that the largest magnitudes of acceleration produced SDH in the parietal lobe, and the lowest in the occipital lobe. Overall this thesis examined the mechanism of injury for TBI using a large dataset with methodologies which complement each other’s limitations. As a result in depth information of the nature of TBI was attained and information provided which may be used to improve future protection and standard development.

Separating unknown signal mixtures into their constituent parts is a difficult problem in signal processing called blind source separation. One of the benchmark problems in this area is the extraction of the fetal heartbeat from an electrocardiogram in which it is overshadowed by a strong maternal heartbeat. This thesis presents a study of a signal separation technique called independent component analysis (ICA), in order to assess its suitability for the maternal-fetal ECG separation problem. This includes an analysis of ICA on deterministic, stochastic, simulated and recorded ECG signals. The experiments presented in this thesis demonstrate that ICA is effective on linear mixtures of known simulated or recorded ECGs. The performance of ICA was measured using visual comparison, heart rate extraction, and energy, information theoretic, and fractal-based measures. ICA extraction of clinically recorded maternal-fetal ECGs mixtures, in which the source signals were unknown, were successful at recovering the fetal heart rate.

This report intends to verify the possibility that the FastICA algorithm could be applied to the GPS system to eliminate the impulsive noise from the receiver end. As the impulsive noise is so unpredictable in its pattern and of great energy level to swallow the signal we need, traditional signal selection methods exhibit no much use in dealing with this problem. Blind Source Separation seems to be a good way to solve this, but most of the other BSS algorithms beside FastICA showed more or less degrees of dependency on the pattern of the noise. In this thesis, the basic mathematic modelling of this advanced algorithm, along with the principles of the commonly used fast independent component analysis (fastICA) based on fixed-point algorithm are discussed. To verify that this method is useful under industrial use environment to remove the impulsive noises from digital BPSK modulated signals, an observation signal mixed with additive impulsive noise is generated and separated by fastICA method. And in the last part of the thesis, the fastICA algorithm is applied to the GPS receiver modeled in the SoftGNSS project and verified to be effective in industrial applications. The results have been analyzed. / 6

In an unsupervised learning problem, one is given an unlabelled dataset and hopes to find some hidden structure; the prototypical example is clustering similar data. Such problems often arise in machine learning and statistics, but also in signal processing, theoretical computer science, and any number of quantitative scientific fields. The distinguishing feature of unsupervised learning is that there are no privileged variables or labels which are particularly informative, and thus the greatest challenge is often to differentiate between what is relevant or irrelevant in any particular dataset or problem. In the course of this thesis, we study a number of problems which span the breadth of unsupervised learning. We make progress in Gaussian mixtures, independent component analysis (where we solve the open problem of underdetermined ICA), and we formulate and solve a feature selection/dimension reduction model. Throughout, our goal is to give finite sample complexity bounds for our algorithms -- these are essentially the strongest type of quantitative bound that one can prove for such algorithms. Some of our algorithmic techniques turn out to be very efficient in practice as well. Our major technical tool is tensor spectral decomposition: tensors are generalisations of matrices, and often allow access to the "fine structure" of data. Thus, they are often the right tools for unravelling the hidden structure in an unsupervised learning setting. However, naive generalisations of matrix algorithms to tensors run into NP-hardness results almost immediately, and thus to solve our problems, we are obliged to develop two new tensor decompositions (with robust analyses) from scratch. Both of these decompositions are polynomial time, and can be viewed as efficient generalisations of PCA extended to tensors.