The test of homogeneity for intraclass correlation coefficients has been one of the active topics in statistical research. Several chi-square tests have been proposed to test the homogeneity of intraclass correlations in the past few decades. The big concern for them is that these methods are seriously biased when sample sizes are not large. In this thesis, data driven approaches are proposed to testing the homogeneity of intraclass correlation coefficients of several populations. Through simulation study, data driven methods have been proved to be less biased and accurate than some commonly used chi-square tests.

Applied Research in System Analysis (ARiSA) is a company specialized in the development of the customer-specific quality models and applied research work. In order to improve the quality of the projects and to reduce maintenance costs, ARiSA developed Quality Monitor (QM) – a web application for quality analysis. QM application has been originally developed as a basic program to enable customers to evaluate the quality of the sources. Therefore, the business logic of the application was simplified and certain limitations were imposed on it, which in its turn leads to a number of issues related to user experience, performance and architecture design. These aspects are important for both application as a product, and for its future promotion. Moreover, this is important for customers, as end users. Main application issues, which were added to the maintenance list are: manual data upload, insufficient server resources to handle long-running and resource consuming operations, no background processing and status reporting, simplistic presentation of analysis results and known usability issues, weak integration between analysis back-ends and front-end. ­­­­­­­­­­­In order to address known issues and to make improvements of the existing limitations, a maintenance phase of QM application is initiated. First of all, it is intended to stabilize current version and improve user experience. It also needed for refactoring and implementation of more efficient data uploads processing in the background. In addition, extended functionality of QM would fulfill customer needs and transform application from the project into a product. Extended functionality includes: automated data upload from different build processes, new data visualizations, and improvement of the current functionality according to customer comments. Maintenance phase of QM application has been successfully completed and master thesis goals are met. Current version is more stable and more responsive from user experience perspective. Data processing is more efficient, and now it is implemented as background analysis with automatic data import. User interface has been updated with visualizations for client-side interaction and progress reporting. The solution has been evaluated and tested in close cooperation with QM application customers. This thesis describes requirements analysis, technology stack with choice rationale and implementation to show maintenance results.

Objectives: The aim of the two exploratory studies presented here, was to investigate expert-novice cognitive performance in the field of dietetic counseling. More specifically, the purpose was to characterize the knowledge used and the cognitive reasoning strategies of expert, intermediate and novice dietitians during their assessment of clinical vignettes of simulated dyslipidemia cases. Background: Since no studies have been conducted on the expert-novice differences in knowledge utilization and reasoning in the field of dietetics, literature from various domains looking at expert-novice decision-making was used to guide the studies presented here. Previous expert-novice research in aspects of health such as counseling and diagnostic reasoning among physicians and nurses has found differences between in the way experts extract and apply knowledge during reasoning. In addition, various studies illustrate an intermediate effect, where generalist performance is somewhat poorer than that of experts and novices. Methods: The verbal protocols of expert (n=4), generalist (n=4), and novice (n=4) dietitians were analyzed, using propositional analysis. Semantic networks were generated, and used to compare reasoning processes to a reference model developed from an existing Dyslipidemia care map by Brauer et al, (2007, 2009). Detailed analysis was conducted on individual networks in an effort to obtain better understanding of cue utilization, concept usage, and overall cohesiveness during reasoning. Results: The results of the first study indicate no statistical differences in reasoning between novices, generalist and experts with regards to recalls and inferences. Interesting findings in the study also suggest that discussions of the terms “dietary fat” and “cholesterol” by individuals in each level of expertise had qualitative differences. This may be reflective of the information provided in the case scenearios to each participating dietitian. Furthermore, contrary to previous studies in expert-novice reasoning, an intermediate effect was not evident. The results of the second study show a statistical difference in data driven (forward) reasoning between experts and novices. There was no statistical difference in hypothesis driven (backward) reasoning between groups. The reasoning networks of experts appear to reveal more concise explanations of important aspects related to dyslipidemia counseling. Reasoning patterns of the expert dietitians appear more coherent, although there was no statistical difference in the length or number of reasoning chains between groups. With previous research focusing on diagnostic reasoning rather than counseling, this finding may be a result of the nature of the underlying task. Conclusion: The studies presented here serve as a basis for future expert-novice research in the field of dietetics. The exploration of individual verbal protocols to identify characteristics of dietitians of various levels of expertise, can provide insight into the way knowledge is used and applied during diet counseling. Subsequent research can focus on randomized sample selection, with case scenarios as a constant, in order to obtain results that can be generalized to the greater dietitian population.

The Swedish health care sector must improve productivity in order to deal with anincreased demand from an aging population with limited resources. In the traditiondriven health care sector, transitioning towards becoming more data-driven has beenidentified as a potential solution. This explorative qualitative case study explores howindividual employees perceive this development at two Swedish hospitals. The resultscomplement theory by presenting propositions that explains drivers and barriers ofthe transition, but also the outcomes of it as perceived by the employees. The studyprimarily concludes that (1) a lack of trust in data and a tradition to base decisions ongut feelings in conjunction with low IT competence make hospital culture a majorobstacle for the transition, and that (2) it is important to understand the employees’perceived outcomes of becoming data-driven as it affects their support of thetransition. The results provide a platform for future research to build on and arevaluable for practitioners as they seek to utilize the drivers and mitigate the barriers.

In this paper, we investigate rescaling of energy features for noise-robust speech recognition. The performance of the speech recognition system will degrade very quickly by the influence of environmental noise. As a result, speech robustness technique has become an important research issue for a long time. However, many studies have pointed out that the impact of speech recognition under the noisy environment is enormous. Therefore, we proposed the data-driven energy features rescaling (DEFR) to adjust the features. The method is divided into three parts, that are voice activity detection (VAD), piecewise log rescaling function and parameter searching algorithm. The purpose is to reduce the difference of noisy and clean speech features. We apply this method on Mel-frequency cepstral coefficients (MFCC) and Teager energy cepstral coefficients (TECC), and we compare the proposed method with mean subtraction (MS) and mean and variance normalization (MVN). We use the Aurora 2.0 and Aurora 3.0 databases to evaluate the performance. From the experimental results, we proved that the proposed method can effectively improve the recognition accuracy.

In this dissertation, the generalized score tests based on weighted estimating equations are proposed for missing covariate data. Their properties, including the effects of nuisance functions on the forms of the test statistics and efficiency of the tests, are investigated. Different versions of the test statistic are properly defined for various parametric and semiparametric settings. Their asymptotic distributions are also derived. It is shown that when models for the nuisance functions are correct, appropriate test statistics can be obtained via plugging the estimates of the nuisance functions into the appropriate test statistic for the case that the nuisance functions are known. Furthermore, the optimal test is obtained using the relative efficiency measure. As an application of the proposed tests, a formal model validation procedure is developed for generalized linear models in the presence of missing covariates. The asymptotic distribution of the data driven methods is provided. A simulation study in both linear and logistic regressions illustrates the applicability and the finite sample performance of the methodology. Our methods are also employed to analyze a coronary artery disease diagnostic dataset.

Thesis (M.A.)--Indiana University, 2009. / Title from screen (viewed on February 1, 2009). Department of English, Indiana University-Purdue University Indianapolis (IUPUI). Advisor(s): Julie A. Belz, Ulla M. Connor, Thomas A. Upton. Includes vitae. Includes bibliographical references (leaves 65-67).

Today’s process control industry, which is extensively automated, generates huge amounts of process data from the sensors used to monitor the processes. These data if effectively analyzed and interpreted can give a clearer picture of the performance of the underlying process and can be used for its proactive monitoring. With the great advancements in computing systems a new genre of process monitoring and fault detection systems are being developed which are essentially data-driven. The objectives of this research are to explore a set of data-driven methodologies with a motive to provide a predictive modeling framework and to apply it to process control. This project explores some of the data-driven methods being used in the process control industry, compares their performance, and introduces a novel method based on statistical process control techniques. To evaluate the performance of this novel predictive modeling technique called Multi-state PLS, a patented continuous process analytics technique that is being developed at Emerson Process Management, Austin, some extensive simulations were performed in MATLAB. A MATLAB Graphical User Interface has been developed for implementing the algorithm on the data generated from the simulation of a continuously stirred blending tank. The effects of noise, disturbances, and different excitations on the performance of this algorithm were studied through these simulations. The simulations have been performed first on a steady state system and then applied to a dynamic system .Based on the results obtained for the dynamic system, some modifications have been done in the algorithm to further improve the prediction performance when the system is in dynamic state. Future work includes implementing of the MATLAB based predictive modeling technique to real production data, assessing the performance of the algorithm and to compare with the performance for simulated data. / text

Retention in Computer Science undergraduate education, particularly of underrepresented groups, continues to be a growing challenge. A theme shared by much of the research literature into why this is so is one of a distancing in the relationship between Computer Science professors and students [39, 40, 45]. How then, can we begin to lessen that distance, and build stronger connections between these groups in an era of growing class sizes and technology replacing human interaction? This work presents BitFit, an online programming practice and learning tool, to describe an approach to using the telemetry made possible from deploying this or similar tools in introductory programming courses to improve the quality of instruction, and the students' course experiences. BitFit gathers interaction data as students use the tool to actively engage with course material. In this thesis we first explore what kind of quantitative data can be used to help professors gain insights into how students might be faring in their courses, moving the method of instruction towards a data- and student-driven model. Secondly, we demonstrate the capacity of the telemetry to aid professors in more precisely identifying students at risk of failure in their courses. Our goal is to reveal possible reasons these students would be considered at-risk at an early enough point in the course to make interventions possible. Finally, we show how the use of tools such as BitFit within introductory programming courses could positively impact the student experience. Through a preliminary qualitative assessment, we seek to address impact on confidence, metacognition, and the ability for an individual to envision success in Computer Science. When used together within an all-encompassing approach aimed at improving retention in Computer Science, tools such as BitFit can move towards improving the quality of instruction and the students' experience by helping to build stronger connections rooted in empathy between professors and students. / Graduate / 0710 / 0984 / alrusso@uvic.ca

Dynamically determining input signals to a complex system, to increase performance and/or reduce cost, is a difficult task unless users are provided with feedback on the consequences of different input decisions. For example, users self-determine the set point schedule (i.e. temperature thresholds) of their HVAC system, without an ability to predict cost--they select only comfort. Users are unable to optimize the set point schedule with respect to cost because the cost feedback is provided at billing-cycle intervals. To provide rapid feedback (such as expected monthly/daily cost), mechanisms for system monitoring, data-driven modeling, simulation, and optimization are needed. Techniques from the literature require in-depth knowledge in the domain, and/or significant investment in infrastructure or equipment to measure state variables, making these solutions difficult to implement or to scale down in cost. This work introduces methods to approximate complex system behavior prediction and optimization, based on dynamic data obtained from inexpensive sensors. Unlike many existing approaches, we do not extract an exact model to capture every detail of the system; rather, we develop an approximated model with key predictive characteristics. Such a model makes estimation and prediction available to users who can then make informed decisions; alternatively, these estimates are made available as an input to an optimization tool to automatically provide pareto-optimized set points. Moreover, the approximation nature of this model makes the determination of the prediction and optimization parameters computationally inexpensive, adaptive to system or environment change, and suitable for embedded system implementation. Effectiveness of these methods is first demonstrated on an HVAC system methodology, and then extended to a variety of complex system applications.