The complex human motor function of speech presents a scientifically interesting, yet relatively unexplored, means to study brain-behavior relationships. Fortunately, magnetic resonance imaging (MRI), which has been proven to characterize soft tissue excellently, has recently become a promising technique for the study of speech. MRI\'s contributions in speech research could lead to new and individualized treatment for speech disorders. Although many studies have shown that MRI can capture information about speech, this project sought to determine what covert information could be disclosed from MRI movies through multivariate analysis. The articulation of phoneme pairs was imaged using a novel sequence, and simultaneously recorded. The data were then analyzed using support vector machine (SVM) analysis and canonical correlation analysis (CCA). Determination of classification accuracy through SVM analysis revealed that phoneme pairs were distinguishable from one another consistently over 90% of the time using information found from MRI movie clips of the speech. Additionally, study of the SVM weights demonstrated that SVM could identify regions of the vocal tract that are used to form auditory distinctions between the phonemes. Finally, CCA revealed relationships between images and the frequencies in corresponding audio waveforms; once again, the speech articulators were identified as lending maximum correlation to the sound profile. These promising results demonstrate that multivariate analysis can uncover information that is known to be true concerning speech production. These analyses may perhaps even contribute to existing knowledge and thus provide a platform from which to advance the treatment of speech dysfunction. / Master of Science

The standard methodology when building statistical models has been to use one of several algorithms to systematically search the model space for a good model. If the number of variables is small then all possible models or best subset procedures may be used, but for data sets with a large number of variables, a stepwise procedure is usually implemented. The stepwise procedure of model selection was designed for its computational efficiency and is not guaranteed to find the best model with respect to any optimality criteria. While the model selected may not be the best possible of those in the model space, commonly it is almost as good as the best model. Many times there will be several models that exist that may be competitors of the best model in terms of the selection criterion, but classical model building dictates that a single model be chosen to the exclusion of all others. An alternative to this is Bayesian model averaging (BMA), which uses the information from all models based on how well each is supported by the data. Using BMA allows a variance component due to the uncertainty of the model selection process to be estimated. The variance of any statistic of interest is conditional on the model selected so if there is model uncertainty then variance estimates should reflect this. BMA methodology can also be used for variable assessment since the probability that a given variable is active is readily obtained from the individual model posterior probabilities. The multivariate methods considered in this research are principal components analysis (PCA), canonical variate analysis (CVA), and canonical correlation analysis (CCA). Each method is viewed as a particular multivariate extension of univariate multiple regression. The marginal likelihood of a univariate multiple regression model has been approximated using the Bayes information criteria (BIC), hence the marginal likelihood for these multivariate extensions also makes use of this approximation. One of the main criticisms of multivariate techniques in general is that they are difficult to interpret. To aid interpretation, BMA methodology is used to assess the contribution of each variable to the methods investigated. A second issue that is addressed is displaying of results of an analysis graphically. The goal here is to effectively convey the germane elements of an analysis when BMA is used in order to obtain a clearer picture of what conclusions should be drawn. Finally, the model uncertainty variance component can be estimated using BMA. The variance due to model uncertainty is ignored when the standard model building tenets are used giving overly optimistic variance estimates. Even though the model attained via standard techniques may be adequate, in general, it would be difficult to argue that the chosen model is in fact the correct model. It seems more appropriate to incorporate the information from all plausible models that are well supported by the data to make decisions and to use variance estimates that account for the uncertainty in the model estimation as well as model selection. / Ph. D.

American democracy fosters the common good of society by allowing citizen involvement in government. Sustaining American democracy depends on civic involvement among citizens. Civic involvement, which consists of citizens' informed involvement in government, politics, and community life, is a desired behavior among adult citizens in the United States and it is a desired outcome of higher education. However, people in the latter part of the twentieth century have questioned the extent to which higher education makes a difference in civic involvement among adults in the United States. College educators are challenged to explain the relationship between higher education and civic involvement among adults in the 1990s. The purpose of the present study is to investigate the relationship between higher education and civic involvement. The researcher approached this issue by examining relationships between measures of civic involvement and personal characteristics such as education level, race, gender, age, and socioeconomic status among adults in the United States. The researcher compared joint relationships between civic involvement and personal characteristics among college graduates with the joint relationships between civic involvement and personal characteristics among adults with some college education and adults with no college education. Data from the Adult Civic Involvement component of the National Household Education Survey of 1996 (NHES:96) were analyzed. This survey was conducted by the National Center for Education Statistics. Using list-assisted, random digit dialing methods and computer assisted telephone interviewing techniques, data were collected from a nationally representative sample of non-institutionalized civilians who were eighteen years of age or older at the time of the survey. Data were collected regarding respondents' (a) personal characteristics, (b) use of information sources, (c) knowledge of government, (d) community participation, and (e) political participation. The selected technique for analyzing data was canonical correlation analysis (CCA), which is a form of multivariate analysis that subsumes multiple regression, multivariate analysis of variance, and discriminant analysis. The results revealed that civic involvement among adults in the United States is moderate at best. Low to moderate civic involvement among adults is mostly attributed to the absence of civic behaviors among adults with no college education. Among adults, overall civic involvement has strong relationships with education level, race, gender, age, and socioeconomic status. While the relationship between higher education and civic involvement is strong, there are significant differences in civic involvement among college graduates when grouped according to race, gender, age, and socioeconomic status. White male college graduates with high incomes tend to demonstrate the attributes of civic involvement to a greater extent than other groups. Among adults with some college education, overall civic involvement is characteristic of older males.Similarly, older adults with no college education demonstrate civic involvement to a greater extent than younger adults with no college education. These findings are consistent with the results of previous studies. The findings also extend the results of previous studies by explaining the relationships between civic involvement and multiple personal characteristics when analyzed simultaneously. The findings suggest a need for ongoing analyses of civic involvement among adult citizens and among college students. The results further imply a need for college personnel to identify and implement strategies that will improve the civic outcomes of higher education for minorities and females in various age and income categories. / Ph. D.

The study was aimed at exploring the relationship between poverty and aggravated robbery in Limpopo Province. Sampled secondary data of aggravated robbery of- fenders, obtained from the South African Police (SAPS), Polokwane, was used in the analysis. From empirical researches on poverty and crime, there are some deductions that vulnerability to crime is increased by poverty. Poverty set was categorised by gender, employment status, marital status, race, age and educational attainment. Variables for aggravated robbery were house robbery, bank robbery, street/common robbery, carjacking, truck hijacking, cash-in-transit and business robbery. Canonical correlation analysis was used to make some inferences about the relationship of these two sets. The results revealed a signi cant positive correlation of 0.219(p-value = 0.025) between poverty and aggravated robbery at ve per cent signi cance level. Of the thirteen variables entered into the poverty-aggravated model, ve emerged as sta- tistically signi cant. These were gender, marital status, employment status, common robbery and business robbery. / Mathematical Sciences / M. Sc. (Statistics)

Feature extraction is an extremely important pre-processing step to pattern recognition, and machine learning problems. This thesis highlights how one can best extract features from the data in an exhaustively online and purely adaptive manner. The solution to this problem is given for both labeled and unlabeled datasets, by presenting a number of novel on-line learning approaches. Specifically, the differential equation method for solving the generalized eigenvalue problem is used to derive a number of novel machine learning and feature extraction algorithms. The incremental eigen-solution method is used to derive a novel incremental extension of linear discriminant analysis (LDA). Further the proposed incremental version is combined with extreme learning machine (ELM) in which the ELM is used as a preprocessor before learning. In this first key contribution, the dynamic random expansion characteristic of ELM is combined with the proposed incremental LDA technique, and shown to offer a significant improvement in maximizing the discrimination between points in two different classes, while minimizing the distance within each class, in comparison with other standard state-of-the-art incremental and batch techniques. In the second contribution, the differential equation method for solving the generalized eigenvalue problem is used to derive a novel state-of-the-art purely incremental version of slow feature analysis (SLA) algorithm, termed the generalized eigenvalue based slow feature analysis (GENEIGSFA) technique. Further the time series expansion of echo state network (ESN) and radial basis functions (EBF) are used as a pre-processor before learning. In addition, the higher order derivatives are used as a smoothing constraint in the output signal. Finally, an online extension of the generalized eigenvalue problem, derived from James Stone’s criterion, is tested, evaluated and compared with the standard batch version of the slow feature analysis technique, to demonstrate its comparative effectiveness. In the third contribution, light-weight extensions of the statistical technique known as canonical correlation analysis (CCA) for both twinned and multiple data streams, are derived by using the same existing method of solving the generalized eigenvalue problem. Further the proposed method is enhanced by maximizing the covariance between data streams while simultaneously maximizing the rate of change of variances within each data stream. A recurrent set of connections used by ESN are used as a pre-processor between the inputs and the canonical projections in order to capture shared temporal information in two or more data streams. A solution to the problem of identifying a low dimensional manifold on a high dimensional dataspace is then presented in an incremental and adaptive manner. Finally, an online locally optimized extension of Laplacian Eigenmaps is derived termed the generalized incremental laplacian eigenmaps technique (GENILE). Apart from exploiting the benefit of the incremental nature of the proposed manifold based dimensionality reduction technique, most of the time the projections produced by this method are shown to produce a better classification accuracy in comparison with standard batch versions of these techniques - on both artificial and real datasets.

Reliability prediction of components operating in complex systems has historically been conducted in a statistically isolated manner. Current physics-based, i.e. mechanistic, component reliability approaches focus more on component-specific attributes and mathematical algorithms and not enough on the influence of the system. The result is that significant error can be introduced into the component reliability assessment process. The objective of this study is the development of a framework that infuses the influence of the system into the process of conducting mechanistic-based component reliability assessments. The formulated framework consists of six primary steps. The first three steps, identification, decomposition, and synthesis, are qualitative in nature and employ system reliability and safety engineering principles for an appropriate starting point for the component reliability assessment. The most unique steps of the framework are the steps used to quantify the system-driven local parameter space and a subsequent step using this information to guide the reduction of the component parameter space. The local statistical space quantification step is accomplished using two newly developed multivariate probability tools: Multi-Response First Order Second Moment and Taylor-Based Inverse Transformation. Where existing joint probability models require preliminary statistical information of the responses, these models combine statistical information of the input parameters with an efficient sampling of the response analyses to produce the multi-response joint probability distribution. Parameter space reduction is accomplished using Approximate Canonical Correlation Analysis (ACCA) employed as a multi-response screening technique. The novelty of this approach is that each individual local parameter and even subsets of parameters representing entire contributing analyses can now be rank ordered with respect to their contribution to not just one response, but the entire vector of component responses simultaneously. The final step of the framework is the actual probabilistic assessment of the component. Variations of this final step are given to allow for the utilization of existing probabilistic methods such as response surface Monte Carlo and Fast Probability Integration. The framework developed in this study is implemented to conduct the finite-element based reliability prediction of a gas turbine airfoil involving several failure responses. The framework, as implemented resulted in a considerable improvement to the accuracy of the part reliability assessment and an increased statistical understanding of the component failure behavior.

The study was aimed at exploring the relationship between poverty and aggravated robbery in Limpopo Province. Sampled secondary data of aggravated robbery of- fenders, obtained from the South African Police (SAPS), Polokwane, was used in the analysis. From empirical researches on poverty and crime, there are some deductions that vulnerability to crime is increased by poverty. Poverty set was categorised by gender, employment status, marital status, race, age and educational attainment. Variables for aggravated robbery were house robbery, bank robbery, street/common robbery, carjacking, truck hijacking, cash-in-transit and business robbery. Canonical correlation analysis was used to make some inferences about the relationship of these two sets. The results revealed a signi cant positive correlation of 0.219(p-value = 0.025) between poverty and aggravated robbery at ve per cent signi cance level. Of the thirteen variables entered into the poverty-aggravated model, ve emerged as sta- tistically signi cant. These were gender, marital status, employment status, common robbery and business robbery. / Mathematical Sciences / M. Sc. (Statistics)

Principal Component Analysis (PCA) and Canonical Correlation Analysis (CCA) are widely-used mathematical models for decomposing multivariate data. They capture spatial relationships between variables, but ignore any temporal relationships that might exist between observations. Probabilistic PCA (PPCA) and Probabilistic CCA (ProbCCA) are versions of these two models that explain the statistical properties of the observed variables as linear mixtures of an alternative, hypothetical set of hidden, or latent, variables and explicitly model noise. Both the noise and the latent variables are assumed to be Gaussian distributed. This thesis introduces two new models, named PPCA-AR and ProbCCA-AR, that augment PPCA and ProbCCA respectively with autoregressive processes over the latent variables to additionally capture temporal relationships between the observations. To make PPCA-AR and ProbCCA-AR robust to outliers and able to model leptokurtic data, the Gaussian assumptions are replaced with infinite scale mixtures of Gaussians, using the Student-t distribution. Bayesian inference calculates posterior probability distributions for each of the parameter variables, from which we obtain a measure of confidence in the inference. It avoids the pitfalls associated with the maximum likelihood method: integrating over all possible values of the parameter variables guards against overfitting. For these new models the integrals required for exact Bayesian inference are intractable; instead a method of approximation, the variational Bayesian approach, is used. This enables the use of automatic relevance determination to estimate the model orders. PPCA-AR and ProbCCA-AR can be viewed as linear dynamical systems, so the forward-backward algorithm, also known as the Baum-Welch algorithm, is used as an efficient method for inferring the posterior distributions of the latent variables. The exact algorithm is tractable because Gaussian assumptions are made regarding the distribution of the latent variables. This thesis introduces a variational Bayesian forward-backward algorithm based on Student-t assumptions. The new models are demonstrated on synthetic datasets and on real remote sensing and EEG data.

Central America is a region regularly affected by natural disasters, with most of them having a hydro-meteorological origin. Therefore, the understanding of annual changes of precipitation upon the region is relevant for planning and mitigation of natural disasters. This thesis focuses on studying the precipitation variability at annual scales in Central America within the framework of the Swedish Centre for Natural Disaster Science. The aims of this thesis are: i) to establish the main climate variability sources during the boreal winter, spring and summer by using different statistical techniques, and ii) to study the connection of sea surface temperature anomalies of the neighbouring oceans with extreme precipitation events in the region. Composites analysis is used to establish the variability sources during winter. Canonical correlation analysis is employed to explore the connection between the SST anomalies and extreme rainfall events during May-June and August-October. In addition, a global circulation model is used to replicate the results found with canonical correlation analysis, but also to study the relationship between the Caribbean Sea surface temperature and the Caribbean low-level jet. The results show that during winter both El Niño Southern Oscillation and the Pacific Decadal Oscillation, are associated with changes of the sea level pressure near the North Atlantic Subtropical High and the Aleutian low. In addition, the El Niño Southern Oscillation signal is intensified (destroyed) when El Niño and the Pacific Decadal Oscillation have the same (opposite) sign. Sea surface temperature anomalies have been related to changes in both the amount and temporal distribution of rainfall. Precipitation anomalies during May-June are associated with sea surface temperature anomalies over the Tropical North Atlantic region. Whereas, precipitation anomalies during August-September-October are associated with the sea surface temperature anomalies contrast between the Pacific Ocean and the Tropical North Atlantic region. Model outputs show no association between sea surface temperature gradients and the Caribbean low-level jet intensification. Canonical correlation analysis shows potential for prediction of extreme precipitation events, however, forecast validation shows that socio-economic variables must be included for more comprehensive natural disaster assessments.

In an information era of new knowledge-based economy, how should the traditional medical industry respond? Do medical professionals have any awareness of knowledge management? If yes, what is their attitude towards knowledge? What are the features? Based on the infusion methodology of knowledge management, if a hospital seeks to adopt an information system for knowledge management (KM), the hospital will have to go through the process of integrating the strategy, the process, the information technology and the awareness of the people and the organization. Subjects of this study were medical professionals, and Arthur Andersen¡¦s knowledge management model was taken as reference. The structure of the study is divided into two sections, i.e. ¡§employees¡¦ awareness of knowledge management¡¨ and ¡§employees¡¦ awareness of the hospital¡¦s knowledge management¡¨. In the former section, factor analysis reveals that medical professionals tend to view knowledge management as the planning and integration of personal knowledge management. Medical professionals recognize that knowledge is an important personal asset, and some inexpressible tacit knowledge still exists in one¡¦s knowledge that needs to be valued. Furthermore, they also believe that in order to increase the value of knowledge assets, it is necessary to share knowledge. In the latter section, factor analysis reveals that medical professionals recognize that a hospital should have a properly planned process to gain knowledge and should create an innovative and practical environment to accumulate knowledge-based assets for the hospital. As for personal property, in the section of ¡§employees¡¦ awareness of knowledge management¡¨, the study manifests obvious difference on the variables of personal property which includes scale of hospital, age, gender, working years, educational background, type of work and management post. On the other hand, the study shows that in the section of ¡§employees¡¦ awareness of the hospitals¡¦ knowledge management¡¨, there is obvious difference on the variables of personal property including scale of hospital, age, working year, educational background while the property of the gender, type of work and management post remain relatively the same. Moreover, the study also shows that if employees are more satisfied with the self-awareness of knowledge management, they are more likely to agree with every dimensions of the knowledge management of the hospital they work in. On the other hand, if all dimensions of the hospital¡¦s knowledge management are recognized by the employees, the variable of the self-awareness of knowledge management reflected on the employees will also be positive. This demonstrates that the demand of employee¡¦s personal awareness will be influenced by the working environment, and vice versa. Outstanding employees will affect the operation performance of the hospital. Similarly, a properly managed hospital will enhance the performance of the employees. It is apparent that the two are related to each other from the feature analysis of the study. Thus, if the hospital can invest in improving the structure of the knowledge assets of the hospital and enhance education and propagation, any improvement of the ability of knowledge management of either employers or employees will definitely benefit the other party. In regards to the four elements including the promotion of knowledge management, high-level leadership enterprise culture, information technology and measurement indicators, the canonical correlation analysis and the regression analysis show that the four elements have been highly approved by the medical professionals. In terms of high-level leadership, the establishment of the Chief Knowledge Office (CKO) has been applauded by the employees. At the meantime, the employees¡¦ ability of using computers and the willingness of sharing information have been greatly influenced by the culture of the enterprise. As for information technology, the information reveals the close relationship between the employees¡¦ ability of using computer and the result of the hospital performance of using the information technology. For measurement indicators, in order to make medical professionals recognize the introduction of knowledge management, performance evaluation of all phases of the introduction of knowledge management should be properly conducted. It is recommended by the study to adopt the theory of the balanced scorecard to be the measurement indicators for evaluation the practical performance of the three components of knowledge assets because the framework of both theories are quite similar. Furthermore, the study reveals that employees who have worked longer and who are older will be more likely to recogize the management of knowledge self-supervision and are more willing to share their knowledge. Apparently, when a hospital tries to introduce knowledge management to the employees, it will be more efficient and more effective if the hospital can make full use of the participation of senior employees.