Methods for Imputing Missing Values and Synthesizing Confidential Values for Continuous and Magnitude Data

Abstract

Continuous variable is one of the major data types collected by the survey organizations. It can be incomplete such that the data collectors need to fill in the missingness. Or, it can contain sensitive information which needs protection from re-identification. One of the approaches to protect continuous microdata is to sum them up according to different cells of features. In this thesis, I represents novel methods of multiple imputation (MI) that can be applied to impute missing values and synthesize confidential values for continuous and magnitude data.

The first method is for limiting the disclosure risk of the continuous microdata whose marginal sums are fixed. The motivation for developing such a method comes from the magnitude tables of non-negative integer values in economic surveys. I present approaches based on a mixture of Poisson distributions to describe the multivariate distribution so that the marginals of the synthetic data are guaranteed to sum to the original totals. At the same time, I present methods for assessing disclosure risks in releasing such synthetic magnitude microdata. The illustration on a survey of manufacturing establishments shows that the disclosure risks are low while the information loss is acceptable.

The second method is for releasing synthetic continuous micro data by a nonstandard MI method. Traditionally, MI fits a model on the confidential values and then generates multiple synthetic datasets from this model. Its disclosure risk tends to be high, especially when the original data contain extreme values. I present a nonstandard MI approach conditioned on the protective intervals. Its basic idea is to estimate the model parameters from these intervals rather than the confidential values. The encouraging results of simple simulation studies suggest the potential of this new approach in limiting the posterior disclosure risk.

The third method is for imputing missing values in continuous and categorical variables. It is extended from a hierarchically coupled mixture model with local dependence. However, the new method separates the variables into non-focused (e.g., almost-fully-observed) and focused (e.g., missing-a-lot) ones. The sub-model structure of focused variables is more complex than that of non-focused ones. At the same time, their cluster indicators are linked together by tensor factorization and the focused continuous variables depend locally on non-focused values. The model properties suggest that moving the strongly associated non-focused variables to the side of focused ones can help to improve estimation accuracy, which is examined by several simulation studies. And this method is applied to data from the American Community Survey.

Extending Science Gateway Frameworks to Support Big Data Applications in the Cloud

Cloud computing offers massive scalability and elasticity required by many scien-tific and commercial applications. Combining the computational and data handling capabilities of clouds with parallel processing also has the potential to tackle Big Data problems efficiently. Science gateway frameworks and workflow systems enable application developers to implement complex applications and make these available for end-users via simple graphical user interfaces. The integration of such frameworks with Big Data processing tools on the cloud opens new oppor-tunities for application developers. This paper investigates how workflow sys-tems and science gateways can be extended with Big Data processing capabilities. A generic approach based on infrastructure aware workflows is suggested and a proof of concept is implemented based on the WS-PGRADE/gUSE science gateway framework and its integration with the Hadoop parallel data processing solution based on the MapReduce paradigm in the cloud. The provided analysis demonstrates that the methods described to integrate Big Data processing with workflows and science gateways work well in different cloud infrastructures and application scenarios, and can be used to create massively parallel applications for scientific analysis of Big Data.

The Open Data Smart City: A Case Study of Smart City Progress through Open Data Initiatives

Smart cities, cities that are supported by an extensive digital infrastructure of sensors, databases and intelligent applications, have become a major area of academic, governmental and public interest. Simultaneously, there has been a growing interest in open data, the unrestricted use of organizational data for public viewing and use. Drawing on Science and Technology Studies (STS), Urban Studies and Political Economy, this thesis examines how digital processes, open data and the physical world can be combined in smart city development, through the qualitative interview-based case study of a Southern Ontario Municipality, Anytown. The thesis asks what are the challenges associated with smart city development and open data proliferation, is open data complimentary to smart urban development; and how is expertise constructed in these fields? The thesis concludes that smart city development in Anytown is a complex process, involving a variety of visions, programs and components. Although smart city and open data initiatives exist in Anytown, and some are even overlapping and complementary, smart city development is in its infancy. However, expert informants remained optimistic, faithful to a technologically sublime vision of what a smart city would bring. The thesis also questions the notion of expertise within the context of smart city and open data projects, concluding that assertions of expertise need to be treated with caution and scepticism when considering how knowledge is received, generated, interpreted and circulates, within organizations.

An R-package for the Estimation and Testing of Multiple Covariates and Biomarker Interactions for Survival Data Based on Local Partial Likelihood

When we study the variables that a ffect survival time, we usually estimate their eff ects by the Cox regression model. In biomedical research, e ffects of the covariates are often modi ed by a biomarker variable. This leads to covariates-biomarker interactions. Here biomarker is an objective measurement of the patient characteristics at baseline. Liu et al. (2015) has built up a local partial likelihood bootstrap model to estimate and test this interaction e ffect of covariates and biomarker, but the R code developed by Liu et al. (2015) can only handle one variable and one interaction term and can not t the model with adjustment to nuisance variables. In this project, we expand the model to allow adjustment to nuisance variables, expand the R code to take any chosen interaction terms, and we set up many parameters for users to customize their research. We also build up an R package called "lplb" to integrate the complex computations into a simple interface. We conduct numerical simulation to show that the new method has excellent fi nite sample properties under both the null and alternative hypothesis. We also applied the method to analyze data from a prostate cancer clinical trial with acid phosphatase (AP) biomarker.

Dislocated Data Palm (in Two Parts)

Portikus presents the work of late British conceptual artist John Latham (1921-2006) and a new work by artist and researcher Neal White. By facilitating a dialogue between these two practices, the show decodes Latham’s expansive and hugely complex oeuvre and the conceptual legacy of art in relation to the ‘event’ as a structural entity.

Predicting the Dispersal Potential of an Invasive Polychaete Pest along a Complex Coastal Biome

Boccardia proboscidea is a recently introduced polychaete in South Africa where it is a notorious pest of commercially reared abalone. Populations were originally restricted to abalone farms but a recent exodus into the wild at some localities has raised conservation concerns due to the species’ invasive status in other parts of the world. Here, we assessed the dispersal potential of B. proboscidea by using a population genetic and oceanographic modeling approach. Since the worm is in its incipient stages of a potential invasion, we used the closely related Polydora hoplura as a proxy due its similar reproductive strategy and its status as a pest of commercially reared oysters in the country. Populations of P. hoplura were sampled from seven different localities and a section of the mtDNA gene, Cyt b and the intron ATPSa was amplified. A high resolution model of the coastal waters around southern Africa was constructed using the Regional Ocean Modeling System. Larvae were represented by passive drifters that were deployed at specific points along the coast and dispersal was quantified after a 12-month integration period. Our results showed discordance between the genetic and modeling data. There was low genetic structure (Φ = 0.04 for both markers) and no geographic patterning of mtDNA and nDNA haplotypes. However, the dispersal model found limited connectivity around Cape Point—a major phylogeographic barrier on the southern African coast. This discordance was attributed to anthropogenic movement of larvae and adult worms due to vectors such as aquaculture and shipping. As such, we hypothesized that cryptic dispersal could be overestimating genetic connectivity. Though wild populations of B. proboscidea could become isolated due to the Cape Point barrier, anthropogenic movement may play the critical role in facilitating the dispersal and spread of this species on the southern African coast.

Predicting the Dispersal Potential of an Invasive Polychaete Pest along a Complex Coastal Biome

Boccardia proboscidea is a recently introduced polychaete in South Africa where it is a notorious pest of commercially reared abalone. Populations were originally restricted to abalone farms but a recent exodus into the wild at some localities has raised conservation concerns due to the species’ invasive status in other parts of the world. Here, we assessed the dispersal potential of B. proboscidea by using a population genetic and oceanographic modeling approach. Since the worm is in its incipient stages of a potential invasion, we used the closely related Polydora hoplura as a proxy due its similar reproductive strategy and its status as a pest of commercially reared oysters in the country. Populations of P. hoplura were sampled from seven different localities and a section of the mtDNA gene, Cyt b and the intron ATPSa was amplified. A high resolution model of the coastal waters around southern Africa was constructed using the Regional Ocean Modeling System. Larvae were represented by passive drifters that were deployed at specific points along the coast and dispersal was quantified after a 12-month integration period. Our results showed discordance between the genetic and modeling data. There was low genetic structure (Φ = 0.04 for both markers) and no geographic patterning of mtDNA and nDNA haplotypes. However, the dispersal model found limited connectivity around Cape Point—a major phylogeographic barrier on the southern African coast. This discordance was attributed to anthropogenic movement of larvae and adult worms due to vectors such as aquaculture and shipping. As such, we hypothesized that cryptic dispersal could be overestimating genetic connectivity. Though wild populations of B. proboscidea could become isolated due to the Cape Point barrier, anthropogenic movement may play the critical role in facilitating the dispersal and spread of this species on the southern African coast.

Human behaviour recognition in data-scarce domains

This paper presents the novel theory for performing multi-agent activity recognition without requiring large training corpora. The reduced need for data means that robust probabilistic recognition can be performed within domains where annotated datasets are traditionally unavailable. Complex human activities are composed from sequences of underlying primitive activities. We do not assume that the exact temporal ordering of primitives is necessary, so can represent complex activity using an unordered bag. Our three-tier architecture comprises low-level video tracking, event analysis and high-level inference. High-level inference is performed using a new, cascading extension of the Rao–Blackwellised Particle Filter. Simulated annealing is used to identify pairs of agents involved in multi-agent activity. We validate our framework using the benchmarked PETS 2006 video surveillance dataset and our own sequences, and achieve a mean recognition F-Score of 0.82. Our approach achieves a mean improvement of 17% over a Hidden Markov Model baseline.

Differential susceptibility of mitochondrial complex II to inhibition by oxaloacetate in brain and heart

Mitochondrial Complex II is a key mitochondrial enzyme connecting the tricarboxylic acid (TCA) cycle and the electron transport chain. Studies of complex II are clinically important since new roles for this enzyme have recently emerged in cell signalling, cancer biology, immune response and neurodegeneration. Oxaloacetate (OAA) is an intermediate of the TCA cycle and at the same time is an inhibitor of complex II with high affinity (Kd ~ 10− 8 M). Whether or not OAA inhibition of complex II is a physiologically relevant process is a significant, but still controversial topic. We found that complex II from mouse heart and brain tissue has similar affinity to OAA and that only a fraction of the enzyme in isolated mitochondrial membranes (30.2 ± 6.0% and 56.4 ± 5.6% in the heart and brain, respectively) is in the free, active form. Since OAA could bind to complex II during isolation, we established a novel approach to deplete OAA in the homogenates at the early stages of isolation. In heart, this treatment significantly increased the fraction of free enzyme, indicating that OAA binds to complex II during isolation. In brain the OAA-depleting system did not significantly change the amount of free enzyme, indicating that a large fraction of complex II is already in the OAA-bound inactive form. Furthermore, short-term ischemia resulted in a dramatic decline of OAA in tissues, but it did not change the amount of free complex II. Our data show that in brain OAA is an endogenous effector of complex II, potentially capable of modulating the activity of the enzyme.

Advanced Business Simulations-Incorporating business and process execution data

Key Performance Indicators (KPIs) and their predictions are widely used by the enterprises for informed decision making. Nevertheless , a very important factor, which is generally overlooked, is that the top level strategic KPIs are actually driven by the operational level business processes. These two domains are, however, mostly segregated and analysed in silos with different Business Intelligence solutions. In this paper, we are proposing an approach for advanced Business Simulations, which converges the two domains by utilising process execution & business data, and concepts from Business Dynamics (BD) and Business Ontologies, to promote better system understanding and detailed KPI predictions. Our approach incorporates the automated creation of Causal Loop Diagrams, thus empowering the analyst to critically examine the complex dependencies hidden in the massive amounts of available enterprise data. We have further evaluated our proposed approach in the context of a retail use-case that involved verification of the automatically generated causal models by a domain expert.

Data-Driven Dynamic Modeling of Coupled Thermal and Electric Outputs of Microturbines

Microturbines are among the most successfully commercialized distributed energy resources, especially when they are used for combined heat and power generation. However, the interrelated thermal and electrical system dynamic behaviors have not been fully investigated. This is technically challenging due to the complex thermo-fluid-mechanical energy conversion processes which introduce multiple time-scale dynamics and strong nonlinearity into the analysis. To tackle this problem, this paper proposes a simplified model which can predict the coupled thermal and electric output dynamics of microturbines. Considering the time-scale difference of various dynamic processes occuring within microturbines, the electromechanical subsystem is treated as a fast quasi-linear process while the thermo-mechanical subsystem is treated as a slow process with high nonlinearity. A three-stage subspace identification method is utilized to capture the dominant dynamics and predict the electric power output. For the thermo-mechanical process, a radial basis function model trained by the particle swarm optimization method is employed to handle the strong nonlinear characteristics. Experimental tests on a Capstone C30 microturbine show that the proposed modeling method can well capture the system dynamics and produce a good prediction of the coupled thermal and electric outputs in various operating modes.

Data-Driven Dynamic Prediction of Interrelated Heat and Electric Outputs of Microturbines

Microturbines are among the most successfully commercialized distributed energy resources, especially when they are used for combined heat and power generation. However, the interrelated thermal and electrical system dynamic behaviors have not been fully investigated. This is technically challenging due to the complex thermo-fluid-mechanical energy conversion processes which introduce multiple time-scale dynamics and strong nonlinearity into the analysis. To tackle this problem, this paper proposes a simplified model which can predict the coupled thermal and electric output dynamics of microturbines. Considering the time-scale difference of various dynamic processes occuring within microturbines, the electromechanical subsystem is treated as a fast quasi-linear process while the thermo-mechanical subsystem is treated as a slow process with high nonlinearity. A three-stage subspace identification method is utilized to capture the dominant dynamics and predict the electric power output. For the thermo-mechanical process, a radial basis function model trained by the particle swarm optimization method is employed to handle the strong nonlinear characteristics. Experimental tests on a Capstone C30 microturbine show that the proposed modeling method can well capture the system dynamics and produce a good prediction of the coupled thermal and electric outputs in various operating modes.

Complex organochlorine pesticide mixtures as determinant factor for breast cancer risk: a population-based case–control study in the Canary Islands (Spain)

[EN]All the relevant risk factors contributing to breast cancer etiology are not fully known. Exposure to organochlorine pesticides has been linked to an increased incidence of the disease, although not all data have been consistent. Most published studies evaluated the exposure to organochlorines individually, ignoring the potential effects exerted by the mixtures of chemicals.

“It’s the People That Drive It”: Motivating and Supporting Collaborative Data Use Practices in Three Teacher Education Programs

Thesis (Ph.D.)--University of Washington, 2016-08

Neutron Scattering Studies on Complex Magnetic Structures in Magnetoelectric Materials

In this thesis, the magnetic properties of four transition-metal oxides are presented. Their multiferroic and magnetoelectric phases have been investigated by means of different neutron scattering techniques. The materials TbMnO3 and MnWO4 belong to the group of spin-induced multiferroics. Their ferroelectric polarization can be explained by the inverse DzyaloshinskiiMoriya interaction. Another common feature of both materials is the presence of subsequent magnetic transitions from a spin-density wave to a spin spiral. The features of the phase transitions have been studied in both materials and it could be shown that diffuse magnetic scattering from the spin spiral is present even in the ordered spin-density wave phase. The excitation spectrum in the multiferroic phase of TbMnO3 was investigated in detail and a comprehensive dataset was obtained using time-of-flight spectroscopy. A spin-wave model could be obtained which can quantitatively describe the full dispersion. Furthermore, the polarization of the zone-center excitations could be derived which fit well to data from inelastic neutron spectroscopy and infrared spectroscopy. With the combination of spherical neutron polarimetry and a poling of the sample by an electric field, it was possible to observe the chiral magnetic component of the magnetic excitations in TbMnO3 and MnWO4. The spin-wave model for TbMnO3 obtained in this thesis is able to correctly describe the dispersion of this component. The double tungstate NaFe(WO4)2 is isostructural to the multiferroic MnWO4 and develops a complex magnetic phase diagram. By the use of neutron diffraction techniques, the zero-field structure and high-field structures in magnetic field applied along the b-axis could be determined. The data reveal a direct transition into an incommensurate spin-spiral structure. The value of the incommensurability is driven by anharmonic modulations and shows strong hysteresis effects. The static and dynamic properties in the magnetoelectric spin-glass phase of Ni0.42Mn0.58TiO3 were studied in detail. The spin-glass phase is composed of short-ranged MnTiO3 and NiTiO3-type order. The antiferromagnetic domains could be controlled by crossed magnetic and electric fields, which was visualized using spherical neutron polarimetry. A comprehensive dataset of the magnetic excitations in the spin-glass phase was collected. The dataset revealed correlations in the hexagonal plane which are only weakly coupled along the c-axis. The excitation spectra could be simulated by taking into account the MnTiO3-type order.