Search Results (1 - 25 of 241 Results)

Sort By  
Sort Dir
 
Results per page  

Menton, WilliamContinuity of Personality Pathology Constructs in an Inpatient Sample: A Comparison of Linear and Count Regression Analyses Using the PID-5 and MMPI-2-RF
MA, Kent State University, 2016, College of Arts and Sciences / Department of Psychology
The present investigation addresses two issues facing contemporary personality assessment researchers – one theoretical, and one methodological. The theoretical issue relates to continuity between categorical models of personality dysfunction (e.g., the polythetic system described in DSM-IV) and emerging dimensional models of pathological personality (e.g., the experimental dimensional model included in DSM-5 for research purposes). Methodologically, the use of linear regression techniques has in recent years been challenged with increasing frequency on the grounds that such techniques make unwarranted (and often incorrect) assumptions about the assessment data they are often used to model. Count regression techniques have been proposed as appropriate modeling alternatives in such cases; however, few attempts have been made to investigate the practical implications of using count regression over linear regression. These theoretical and methodological issues are explored using an inpatient psychiatric sample of participants with suspected personality psychopathology. Personality disorder symptom counts are modeled using both linear and count regression techniques. As predictors in these models, select dimensional personality indicators are drawn from two conceptually similar self-report scale sets – the five broad trait domains from the Personality Inventory for DSM-5 (PID-5), and the PSY-5 scales from the Minnesota Multiphasic Personality Inventory-2-Restructured Form (MMPI-2-RF). Results suggest a high degree of continuity between dimensional and categorical models of personality dysfunction, with PID-5 and PSY-5 scales generally predicting a meaningful amount of variance in personality disorder symptom counts. Moreover, these predictors evidence a differential pattern of association with symptoms of particular categorical personality disorders, suggesting that personality pathology may be specified using dimensional personality traits rather than categorical labels. Furthermore, comparisons between linear and count models suggest that both techniques produce similar results in these data in terms of predictors identified as statistically important and also in terms of relative strengths of trait predictors.

Committee:

Yossef Ben-Porath (Advisor); Manfred van Dulmen (Committee Member); Mary Beth Spitznagel (Committee Member); John Gunstad (Committee Member)

Subjects:

Psychological Tests; Psychology

Keywords:

personality; personality assessment; personality disorders; MMPI; MMPI-2-RF; PID-5; DSM-5; multiple regression; linear regression; count regression; negative binomial regression; self-report inventories

Taslim, CennyMulti-Stage Experimental Planning and Analysis for Forward-Inverse Regression Applied to Genetic Network Modeling
Doctor of Philosophy, The Ohio State University, 2008, Industrial and Systems Engineering
This dissertation proposes methods for steady state linear system identification for both forward cases in which prediction of outputs for new inputs are desired and also inverse prediction of which inputs fostered measured outputs are needed. Special attention is given to genetic network modeling applications. Inverse prediction matters here because then one can predict the effective genetic perturbation associated with a new target drug compound or therapy. The primary application addressed in this dissertation is motivated by our on-going contributions related to Down syndrome which affects approximately 1 out of every 800 children. First, single shot experimentation and analysis to develop network models is considered. The discussion focuses on linear models because of the relevance of equilibrium conditions and the typical scarcity of perturbation data. Yet, deviations from linear systems modeling assumptions are also considered. For system identification, we propose forward network identification regression (FNIR) and experimental planning involving simultaneously perturbing more than a single gene concentration using D-optimal designs. The proposed methods are compared with alternatives using simulation and data sets motivated by the SOS pathway for Escherichia coli bacteria. Findings include that the optimal experimental planning can improve the sensitivity, specificity, and efficiency of the process of deriving genetic networks. In addition, topics for further research are suggested including the need to develop more numerically stable analysis methods, improved diagnostic procedures, sequential design and analysis procedures. Next, multi-stage design and analysis procedures are proposed for experimentation in which both forward and inverse predictions are relevant. Methods are proposed to derive desirable experimental plans for the next batch of tests based on both space filling and D-optimality. The space filling designs are intended to support both linear and nonlinear modeling while D-optimality methods are relatively model-dependent. Rigorous results related to linear optimality criteria are presented in relation to multi-criteria formulations of the forward-inverse problem. Computational results are presented based on the SOS pathway and inspired by an on-going study of the genetic network associated with Down syndrome. In the studied cases, the biologists added a multiple choice constraint to the formulation for their simplicity.

Committee:

Theodore Allen, PhD (Committee Chair); Mario Lauria, PhD (Committee Co-Chair); Clark Mount-Campbell, PhD (Committee Member); Hakan Ferhatosmanoglu, PhD (Committee Member)

Subjects:

Bioinformatics; Biostatistics; Engineering; Operations Research; Statistics

Keywords:

Optimal Design of Experiments; D-Optimality; Inverse Regression; Transcriptional Networks; Statistical Simulation; System Identification; Steady-State; Forward-Inverse Modeling; Bayesian regression

McElyea, RyanTHE IMPACT OF OPPORTUNITY, PROPENSITY, AND DISTAL FACTORS ON SECONDARY EDUCATION SCIENCE, TECHNOLOGY, ENGINEERING, AND MATH (STEM) PROGRAM AND ACADEMIC OUTCOMES
PHD, Kent State University, 2016, College and Graduate School of Education, Health and Human Services / School of Foundations, Leadership and Administration
The 2012 Program for International Student Assessment (PISA) placed the U.S. in the bottom fourth of mathematics achievement, and less than 9% of U.S. 15-year olds were top performers in the same subject. Research into addressing this issue has involved Inquiry Based (IB) programs, such as Project Lead the Way (PLTW). The studies have focused on general Science, Mathematics, Pre-Engineering, state-wide scores, or national assessment scores. Important variables such as individual transcript data, End of Course (EoC) assessment scores, mathematics and/or science Grade Point Average (GPA), or participation in the Biomedicine program of PLTW have not been researched in the context of PLTW programs. Additionally, there is a lack of research using more sophisticated statistical analyses to examine the above relationships. Therefore, the goal of the current study is to determine the relationship between the opportunity factors (i.e., mathematics and science coursework and PLTW coursework), distal factors (i.e., demographics and prior achievement) and propensity factors (i.e., GPA, mathematics and science grades and PLTW grade) with immediate academic year achievement (i.e., EoC scores) with different statistical modeling techniques. Secondly, repeated measures analyses were also used to examine the relationship between the aforementioned variables and academic achievement over time. The Freshmen Model (N = 259) and the Junior Model (N = 73) were developed using Path Analysis. The Sophomore Model (N = 135) and the Senior Model (N = 51) were developed using Hierarchical Multiple Regression. The impact on STEM PLTW grades over time and academic achievement over time (EoC scores) was analyzed by using Repeated Measures Split-Plot ANOVAs and One-Way Repeated Measures ANCOVAs. This exploratory investigation focused on the following main goals: (1) Investigating if a combination of distal, opportunity, and propensity variables can be used to predict current high school year academic achievement, and (2) Determining if a combination of distal, opportunity, and propensity factors can be used to predict high school academic achievement over time (i.e., across the four years of high school from Freshmen to Senior Year). In summary, a few themes emerged from the results of the study. As shown in the Freshmen Model, Gender plays a positive role on the EoC, but a substantial negative role in the Sophomore Model. By the Junior and Senior Models, it plays no role in academic outcome. As shown in examining PLTWG and EoC over time, there is a significant interaction with Gender. These results suggest that PLTW may have a positive effect on females, as EoC and PLTWG improve over time, which may be tied to an increased interest in the STEM fields. Secondly, as mentioned in earlier research, the performance gap between males and females is largely erased, as shown in EoC and PLTWG over time, and the lack of Gender in the Junior Model or Senior Model. By increasing female performance over time and showing that gender plays a non-significant role in predicting academic performance, it would appear that PLTW is moving towards the goal of positively impacting females in STEM (PLTW, 2012). Another trend in the analysis was the relationship between Propensity factors and predicting end of year performance. Math Grade (MG), GPA, and PLTWG were prevalent in three of the four models. By the definition, Propensity factors are impacted by efficacy, effort, and student ability, therefore it may be that this why student achievement is greatly affected by Propensity factors. Also, these variables occurred closely in time to the academic outcomes, which may explain the numerous occurrences in the static models and also the magnitude of the variable coefficients.

Committee:

Aryn Karpinski, PhD (Committee Chair); Tricia Niesz, PhD (Committee Member); Rajeev Rajaram, PhD (Committee Member)

Subjects:

Educational Tests and Measurements; Statistics

Keywords:

STEM; PLTW; Science; Math; Regression; Project Lead the Way; Hiearchical Multiple Regression; Path Analysis; Inquiry-Based Learning; Project Based Learning

Gummadi, JayaramA Comparison of Various Interpolation Techniques for Modeling and Estimation of Radon Concentrations in Ohio
Master of Science in Engineering, University of Toledo, 2013, Engineering (Computer Science)
Radon-222 and its parent Radium-226 are naturally occurring radioactive decay products of Uranium-238. The US Environmental Protection Agency (USEPA) attributes about 10 percent of lung cancer cases that is `around 21,000 deaths per year’ in the United States, caused due to indoor radon. The USEPA has categorized Ohio as a Zone 1 state (i.e. the average indoor radon screening level greater than 4 picocuries per liter). In order to implement preventive measures, it is necessary to know radon concentration levels in all the zip codes of a geographic area. However, it is not possible to survey all the zip codes, owing to reasons such as inapproachability. In such places where radon data are unavailable, several interpolation techniques are used to estimate the radon concentrations. This thesis presents a comparison between recently developed interpolation techniques to new techniques such as Support Vector Regression (SVR), and Random Forest Regression (RFR). Recently developed interpolation techniques include Artificial Neural Network (ANN), Knowledge Based Neural Networks (KBNN), Correction-Based Artificial Neural Networks (CBNN) and the conventional interpolation techniques such as Kriging, Local Polynomial Interpolation (LPI), Global Polynomial Interpolation (GPI) and Radial Basis Function (RBF) using the K-fold cross validation method.

Committee:

William Acosta (Committee Chair); Vijay Devabhaktuni (Committee Co-Chair); Ashok Kumar (Committee Member); Rob Green (Committee Member)

Subjects:

Computer Science

Keywords:

artificial neural networks; cross-validation; correction based artificial neural networks; prior knowledge input; source difference; space-mapped neural networks; support vector regression; radon; random forest regression

Drummer, Talea R.GETTING IN THE GAME: A QUANTITATIVE STUDY OF SECOND-YEAR STUDENT-ATHLETES’ EXPERIENCES UTILIZING EXISTING DATA OF THE 2010 SOPHOMORE EXPERIENCES NATIONAL SURVEY
PHD, Kent State University, 2014, College and Graduate School of Education, Health and Human Services / School of Foundations, Leadership and Administration
The National Collegiate Athletic Association (NCAA) has a variety of rules and regulations that hold intercollegiate athletic departments, teams, and student-athletes accountable to the academic progression of student-athletes. Through various rules and regulations athletes must focus on academic as well as athletic responsibilities. In an era of increased Academic Progress Rate (APR) minimums and amplified penalties to teams that do not meet those minimums, it is imperative to focus on the student-athlete and find ways for athletic academic administrators, coaches, faculty, and other student-affairs personnel to support their athletes. This study utilized quantitative methods to analyze existing data of the 2010 Sophomore Experiences National Survey to examine the second-year athlete respondents (N = 376) as well as non-athlete second-year students. The methods utilized in this study included Exploratory Factor Analysis, Hierarchical Multiple Regression Analysis, Multiple Regression Analysis, and a Comparison of Correlation Coefficients. The findings of this study suggest that second-year athletes and non-athletes need a connection to campus in order to be certain of their major and intend to re-enroll. The findings also suggest that various areas of satisfaction, goal setting, and managing difficulties can have an affect as well. Finally, athletes and non-athletes were not different on what affected how certain they are of their major; however, there were a few differences in the intent to re-enroll between athletes and non-athletes. The goal is for those who work directly with student-athletes will find ways to implement the findings and suggestions of the research to support this unique sub-population.

Committee:

Stephen Thomas, EdD (Committee Chair); Mark Kretovics, PhD (Committee Member); Jason Schenker, PhD (Committee Member); Kulics Jennifer, PhD (Committee Member)

Subjects:

Academic Guidance Counseling; Education; Higher Education; Higher Education Administration; Sports Management

Keywords:

Athlete; Sophomore Slump; Eligibility; Major Selection; Retention; Academic Progress Rate; APR; Intercollegiate Athletics; Progress Towards Degree; College Student; Factor Analysis; Quantitative; Hierarchical Regression; Multiple Regression; NCAA

Khajuria, SaketA Model to Predict Student Matriculation from Admissions Data
Master of Science (MS), Ohio University, 2007, Industrial and Manufacturing Systems Engineering (Engineering)

Enrollment in a university can be increased by changing any of the following three attributes: the applicant pool size, the marketing strategies for applicants, or the admission standards. Universities are using different predictive modeling techniques to increase enrollment, given their changing demographics, tuition rates, and other factors. This document presents a model for predicting the likelihood that a specific undergraduate applicant will matriculate if admitted. A regression model was built using data on the applicants for the 2004 freshman class. Using this model, applicants from 2005 were evaluated, and matriculation was predicted. Applicants predicted to matriculate did so 47.91% of the time, while students projected not to matriculate only matriculated 28.05% of the time. Considering all four possible outcomes (correct matriculation prediction to incorrect non-matriculation decision), the overall accuracy was 60.2%. The accuracy of the model was similar to studies in the literature.

Committee:

David Koonce (Advisor)

Subjects:

Engineering, Industrial

Keywords:

Predictive Enrollment Modelling; Linear Regression; Logistic Regression

Bandreddy, Neel KamalEstimation of Unmeasured Radon Concentrations in Ohio Using Quantile Regression Forest
Master of Science, University of Toledo, 2014, College of Engineering
The most stable isotope of radon is Radon-222, which is a decay product of radium-226 and an indirect decay product of uranium-238, a natural radioactive element. According to the United States Environmental Protection Agency (USEPA), radon is the primary cause of lung cancer among non-smokers. The USEPA classifies Ohio as a zone 1 state because the average radon screening level is more than 4 picocuries per liter. To perform preventive measures, knowing radon concentration levels in all the zip codes of a geographic area is necessary. However, it is impractical to collect the information from all the zip codes due to its inapproachability. Several interpolation techniques have been implemented by researchers to predict the radon concentrations in places where radon data is not available. Hence, to improve the prediction accuracy of radon concentrations, a new technique called Quantile Regression Forests (QRF) is proposed in this thesis. The conventional techniques like Kriging, Local Polynomial Interpolation (LPI), Global Polynomial Interpolation (GPI), and Radial Basis Function (RBF) estimate output using complex mathematics. Artificial Neural Networks (ANN) have been introduced to overcome this problem. Although ANNs show better prediction accuracy in comparison to more conventional techniques, many issues arise, including local minimization and over fitting. To overcome the inadequacies of existing methods, statistical learning techniques such as Support Vector Regression (SVR) and Random Forest Regression (RFR) were implemented. In this thesis, Quantile Regression Forest (QRF) is introduced and compared with SVR, RFR, and other interpolation techniques using available operational performance measures. The study shows that QRF has least validation error compared with other interpolation techniques.

Committee:

Vijay Devabhaktuni (Committee Chair); Ashok Kumar (Committee Member); Mansoor Alam (Committee Member)

Subjects:

Applied Mathematics; Electrical Engineering; Mathematics

Keywords:

Radon; Kriging; Local Polynomial Interpolation; Global Polynomial Interpolation; Radial Basis Function; Artificial Neural Networks; Random Forest Regression; Quantile Regression Forest; operational performance measures

Thomas, Philip S.A Reinforcement Learning Controller for Functional Electrical Stimulation of a Human Arm
Master of Sciences, Case Western Reserve University, 2009, EECS - Computer and Information Sciences
This thesis demonstrates the feasibility of using reinforcement learning (RL) for functional electrical stimulation (FES) control of a human arm as an improvement over (i) previous closed-loop controllers for upper extremities that are unable to adapt to changing system dynamics and (ii) previous RL controllers that required thousands of arm movements to learn. We describe the relevance of the control task and how it can be applied to help people with spinal cord injuries. We also provide simulations that show previous closed-loop controllers are insufficient. We provide background on possible RL techniques for control, focusing on a continuous actor-critic architecture that uses function approximators for its mappings. We test various function approximators, including Artificial Neural Networks (ANNs) and Locally Weighted Regression (LWR) for this purpose. Next, we introduce a novel function approximator, Incremental Locally Weighted Regression (ILWR), which is particularly suited for use in our RL architecture. We then design, implement, and perform clinically relevant tests using ANNs for the two mappings in the continuous actor-critic. During these trials, unexpected behavior is observed and eventually used to create a hybrid controller (that switches among different learning parameter sets) that can both adapt to changes in arm dynamics in 200 to 300 arm movements and remain stable in the long-term. A non-switching controller with similar performance is achieved using ILWR in place of an ANN for the controller's critic mapping.

Committee:

Michael Branicky, PhD (Advisor); Antonie van den Bogert, PhD (Committee Member); Soumya Ray, PhD (Committee Member)

Subjects:

Computer Science

Keywords:

reinforcement learning; locally wieghted regression; incremental locally weighted regression; LWR; ILWR; ANN; actor-critic; continuous actor-critic; functional electrical stimulation; FES; adaptive control

Vlack, Yvette A.A Diffuse Spectral Reflectance Library of Clay Minerals and Clay Mixtures within the VIS/NIR Bands
MS, Kent State University, 2008, College of Arts and Sciences / Department of Geology
The versatility of diffuse spectral reflectance (DSR) was investigated as a complementary methodology to XRD and XRF when studying clay minerals in stratigraphic sequences. The Analytical Spectral Device (ASD) LabSpec® Pro FR UV/VIS/nIR spectrometer provides an innovative nondestructive methodology that is cost effective, portable, quick, and easy to use with samples in the lab or field. LabSpec® Pro FR spectrometer and similar equipment are remarkable research tools underutilized in the area of clay mixtures. This study develops a new methodology that demonstrates the versatility of the LabSpec® Pro FR and the use of DSR as a tool for generating a spectral library and then determining clay mineralogy of various core samples. Samples from two sources were evaluated: (1) sediment from core MNK3, from a slack water Pleistocene lake near St. Louis, in which stratigraphic changes in clay mineralogy occur down core, and (2) the Ordovician Millbrig K-bentonite (samples from AL, GA, KY, TN, and VA), an altered tephra in which the changes occur laterally in a single horizon. DSR spectral data is validated against XRD, ICP-MS, and XRF data. This spectral library was generated from four primary clays and clay mixtures, consisting over 231 two variable mixtures in 5% increments, by weighted percents and is augmented with spectra from the USGS spectral library. Clay mineral standards were obtained from the Clay Mineral Repository and Wards Natural Science. The aim is to close the gap that currently exists for an expanded spectral library of clay mixtures and explore the DSR variability of clay mixtures. PCA (Principal Component Analysis) was used to correlate the spectral data of the library with the two MNK3 and Millbrig sample sets. Stepwise Linear Regression (SLR) analysis was used with the composite library as an identification tool. By combining PCA analysis of unknowns with SLR against our clay mixture library, we identify our components in an objective, quantifiable way. The model predictors from the analysis gave highly significant R-squared values for the extracted PCA assemblages depending on component. One of the challenges was comparing the XRD clay percents against the predicted models. Frequently, the primary clay was predicted, but not the secondary clay. Basically, the result is an ordinal distribution of the amounts of minerals present in the mixture. Ordinal distributions, as non-parametric data, do not allow the computation of averages or proportions, but tell only relative amounts such as greater, greatest, and least. This may be because both cores represent a four component clay mixtures plus ancillary minerals, as opposed to the two component library. Predictability difficulties may also have been due to confounding factors such as the presence of iron-bearing minerals in the mixture causing what is termed by Balsam, 1999, as the ‘matrix effect’; Balsam also states that iron-bearing minerals such as hematite may be masked by illite and chlorite. The spectral clay mineral library is useful and the methodology pursued has proven successful. However, at this time there is no consistency in the predictability of the data. As a result, future research needs to eliminate intervening factors sequentially to determine various iron components and their impact on readings (Balsam et al, 1999).

Committee:

Joseph Ortiz, PhD (Committee Chair); John Haynes, PhD (Committee Member); Ernest Carlson, PhD (Committee Member); Nancy Grant, PhD (Other)

Subjects:

Geology; Soil Sciences

Keywords:

DSR; diffuse spectral reflectance; clays; clay minerals; clay mixtures; spectral library; PCA; priniciple component ananlysis; stepwise linear regression; SLR; regression models; MNK3; Millbrig; iron-bearing minerals; ASD; Analytical Spectral Device

Diop, LamineAssessing and predicting stream-flow at different time scales in the context of climate change: Case of the upper Senegal River basin
Doctor of Philosophy, The Ohio State University, 2017, Food, Agricultural and Biological Engineering
The objectives of this research were to: (i) investigate different aspects of long-term trend detection in monthly, seasonal, and annual streamflow; (ii) assess climate change impacts on the upper Senegal River basin stream-flow; and (iii) evaluate data driven models to forecast daily stream at the upper Senegal River basin. A preliminary study investigated long-term trends on stream-flow at the upper Senegal River basin. Results showed a trend of decreasing annual stream flow; but this was not significant at p < 0.05 for the whole period. However, when integrating various temporal breaking points, a decreasing trend was significant before the first breaking point (1976) but was reversed for the period of 1976 to 1993. For the monthly series, all months exhibit a non-significant decreasing trend except for the month of June that had an increasing trend. The seasonal series showed a decreasing trend that was significant (p <0.05) at MAMJ season. The extremely low and high daily streamflow had significant positive and negative trends, respectively. This research used five General Circulation Models (CNRM, CSIRO, HadGEM2-CC, HadGEM2-ES, and MIROC5), two RCP (RCP4.5 and RCP8.5) scenarios, and the GR4J hydrological model to evaluate the impact of climate change on stream flow in the near future. The results showed that for all models, there were simulated increases in mean monthly temperature under the RCP 4.5 and RCP 8.5. Increases of temperature ranged from 0.54° C to 2.32° C under RCP 4.5 scenario and from 1.12° C to 2.78° C under RCP8.5. However, models were not consistently in agreement in the direction and magnitude of future precipitation changes for monthly rainfall. Some models predicted an increase while other a decrease. The multi-model ensemble projected a decrease of rainfall for all months except for September. The greatest precipitation increases were in September, at 15.6 mm and 10.1 mm with RCP4.5 and RCP8.5, respectively. The greatest precipitation decreases were between June to August for both scenarios with a maximum in July under RCP 4.5 (-9.53 mm) and RCP 8.5 (-20.1mm). Compared to the 1971 – 2000 reference period, results showed a decrease in annual stream-flow of 2.9% and 7.7% under RCP4.5 and RCP8.5 scenarios, respectively. Monthly variations showed a decrease of stream-flow in wet season for both scenarios. This research verified the utility of the support vector regression (SVR) model and generalized regression neural network (GRNN) to predict one day ahead, the daily river flow in the upper Senegal River basin at the Bafing Makana station in West Africa. Two modeling scenarios were considered: (A): where only stream-flow data were used as an input via antecedent values and (B): where rainfall, evapotranspiration and stream-flow data are used. The results showed that the accuracy of the models varied by scenario. Combining the stream-flow data with rainfall and evapotranspiration can substantially improve the accuracy of the two models to predict one-day ahead stream-flow. A comparison of the optimal SVR and the GRNN models indicated that SVR model had superior performance compared to the GRNN.

Committee:

Larry C Brown, PhD (Advisor)

Subjects:

Agricultural Engineering

Keywords:

climate change, RCP, Support vector regression, generalized regression neural networks, trend test, change point, Senegal River basin

Pax, Benjamin MPrediction of Bronchopulmonary Dysplasia by a Priori and Longitudinal Risk Factors in Extremely Premature Infants
Master of Sciences, Case Western Reserve University, 2018, EECS - Electrical Engineering
Prediction of Bronchopulmonary Dysplasia by a Priori and Longitudinal Risk Factors in Extremely Premature Infants ABSTRACT By BENJAMIN PAX This thesis seeks to identify at risk premature infants for Bronchopulmonary Dysplasia (BPD) and to identify new longitudinal risk factors for BPD by postnatal day. A calculator tool has been created to estimate risk of BPD based on a priori risk factors as well as additional clinical variables. The estimator seeks to estimate outcome based on BPD severity (no BPD, mild, moderate, and severe) as defined by the NIH. The variables examined were as follows: birth weight (BW), gestational age (GA), race and ethnicity, sex, type of respiratory support (no support, nasal cannula/hood, CPAP, ventilator, oscillator), mean airway pressure, supplementary oxygen, oxygen severity index, intermittent hypoxemic events, FiO2, SpO2, Snap score, sepsis, and multiple births. Models were evaluated by using a C statistic on a series of multinomial logistic regression models on postnatal days with C Statistic 0.789 on day 1 to 0.945 on day 56.

Committee:

Kenneth Loparo (Advisor); Juliann Di Fiore (Committee Member); Thomas Raffay (Committee Member); Curtis Tatsuoka (Committee Member)

Subjects:

Computer Science; Electrical Engineering; Epidemiology; Mathematics; Medicine; Statistics

Keywords:

Bronchopulmonary Dysplasia; VLBWI; ELGANs; Multinomial Logistic Regression; Ordinal Logistic Regression; a Priori; Longitudinal; Risk Factors; Prediction; Statistical Model; BPD; ROP

Ghosh, Susmit KumarCharacterizing the performance of a biological analogue to a digital inverter
MS, University of Cincinnati, 2010, Engineering and Applied Science: Electrical Engineering
The process of gene regulation in Phage Lambda, a virus infecting E Coli, is subject to attention for the development of a biological analogue to a digital inverter. We study in detail a model for a bioinverter based on the reactions given by Weiss. The Phage Lambda virus exhibits two phases, lysogeny and lysis. These two phases can be viewed as two stages of a digital inverter. We consider two models of the behavior of the system, i.e., deterministic and stochastic. The effects of various parameters of both the models are studied. We use statistical analysis to maximize the gain of the bioinverter. We determine that gain is attributed to a minimum number of reaction constants and to the initial concentration of protein species. This simplifies the task of the engineer who wishes to optimize the system for various applications.

Committee:

Carla Purdy, C, PhD (Committee Chair); George Purdy, PhD (Committee Member); Wen Ben Jone, PhD (Committee Member)

Subjects:

Computer Science

Keywords:

regression table; Cro; regression; StatTools; K7; Phage Lambda; behavior of the system

McCormick, Carmen Amanda McCaneTEMPORAL SIGNIFICANCE OF MEAN SHERD THICKNESS IN SAN FRANCISCO MOUNTAIN GRAY WARE
MA, University of Cincinnati, 2007, Arts and Sciences : Anthropology
In the past, San Francisco Mountain Gray Ware (SFMGW) bearing archaeological sites in north-central Arizona have been dated using the ceramic-type method. However, many sites do not contain abundant amounts of diagnostic pottery types and consequently cannot be dated by the ceramic-type method. Additionally, the dates applied to sites containing SFMGW are also ambiguous because of the long production ranges assigned to each SFMGW type (i.e., 200 years or longer). To help solve these problems in northern Arizona archaeology, a dating method was developed by Daniel Sorrell that correlates the average thickness of SFMGW assemblages with tree-ring dates; simple regression analyses serve as the basis for Sorrell’s chronometric dating model using independent samples and tree-ring dates. This thesis is a direct test of Sorrell’s dating technique. Results confirm that the thickness of unpainted SFMGW changes more or less predictably through time and can be used to date unexcavated and excavated sites in northern Arizona.

Committee:

Alan Sullivan (Advisor)

Subjects:

Anthropology, Archaeology

Keywords:

San Francisco Mountain Gray Ware; Deadmans Gray; Deadmans Fugitive Red; Deadmans Brown; Linear Regression Analysis; Quadratic Regression Analysis; Upper Basin Archaeological Research Project; Kaibab National Forest

Wang, TianyiTrajectory Similarity Based Prediction for Remaining Useful Life Estimation
PhD, University of Cincinnati, 2010, Engineering and Applied Science: Industrial Engineering

The degradation process of a complex system may be affected by many unknown factors, such as unidentified fault modes, unmeasured operational conditions, engineering variance, environmental conditions, etc. These unknown factors not only complicate the degradation behaviors of the system, but also lower the quality of the collected data for modeling. Due to lack of knowledge and incomplete measurements, certain important context information (e.g. fault modes, operational conditions) of the collected data will be missing. Therefore historical data of the system with a large variety of degradation patterns will be mixed together. With such data, learning a global model for Remaining Useful Life (RUL) prediction becomes extremely hard. This has led us to look for advanced RUL prediction techniques beyond the traditional global models.

In this thesis, a novel RUL prediction method inspired by the Instance Based Learning methodology, called Trajectory Similarity Based Prediction (TSBP), is proposed. In TSBP, the historical instances of a system with life-time condition data and known failure time are used to create a library of degradation models. For a test instance of the same system whose RUL is to be estimated, similarity between it and each of the degradation models is evaluated by computing the minimal weighted Euclidean distance defined on two degradation trajectories. Based on the known failure time, each of the degradation models will produce one RUL estimate for the test instance. The final RUL estimate can then be obtained by aggregating the multiple RUL estimates using a density estimation method.

A case study using the turbofan engine degradation simulation data supplied by NASA Ames is provided to study the performance of TSBP. In this study, the TSBP method has demonstrated significant improvement in performance over a Neural Network based prediction method.

Committee:

Jay Lee, PhD (Committee Chair); Hongdao Huang, PhD (Committee Member); Ernest Hall, PhD (Committee Member); Kai Goebel, PhD (Committee Member)

Subjects:

Mechanical Engineering

Keywords:

Prognostics and Health Management;Remaining Useful Life;Instance Based Learning;Kernel Regression;Kernel Density Estimation;Radial Basis Function

Mitsova-Boneva, DianaModeling the Impact of Land Cover Change on Non-point Source Nitrogen Inputs to Streams at a Watershed Level: Implications for Regional Planning
PhD, University of Cincinnati, 2008, Design, Architecture, Art and Planning : Regional Development Planning

The objective of this research is to assess the impact of future land cover changes on nutrient enrichment of streams. It applies cellular automata (CA) Markov chain model to simulate future land cover change and a GIS-based distributed cell-based model to predict non-point source nitrogen loadings to streams. The integration of the two models provides site-specific information on how the spatial location and extent of urban development can affect nitrogen pollution under dry, normal and wet conditions.

Two scenarios of land cover change, in particular, were examined. The baseline scenario (Scenario 1) involved only minor protection of environmentally sensitive areas. The open space conservation network scenario (Scenario 2) incorporated the principles of "green" infrastructure as outlined by the relevant literature. Scenario 2 was based on protection of riparian areas, floodplains, wetlands, urban open space, and areas with exceedingly shallow depth to seasonally high water table and bedrock. Increased setbacks, where appropriate, were considered. The impact of the projected land cover change under different development scenarios was then examined in terms of nitrogen delivery ratios, total loads and contributing areas. A spatial hydrological model of the watershed was developed under dry, normal and wet conditions. A non-linear regression model was applied to estimate nitrogen trapping efficiencies and delivery ratios based on field characteristics such as slope, saturated hydraulic conductivity, soil mean particle diameter, Manning's roughness coefficient and length of flow. An attenuation factor taking into account cost distance to streams and decay constant was also incorporated into the model to account for transmission losses. Contributing areas of nitrogen delivery to streams were delineated based on the model results.

Committee:

Xinhao Wang, PhD (Committee Chair); David Edelman, PhD (Committee Member); Jan Fritz, PhD (Committee Member); William Shuster, PhD (Committee Member)

Subjects:

Urban Planning

Keywords:

land cover change; cellular automata; TN loading model; non-linear regression; open space conservation network

Ijla, AkramThe Impact of Local Historical Designation on Residential Property Value: An Analysis of Three Slow-Growth and Three Fast-Growth Central Cities in the United States
Doctor of Philosophy in Urban Studies and Public Affairs, Cleveland State University, 2008, Levin College of Urban Affairs

THE IMPACT OF LOCAL HISTORIC DESIGNATION ON RESIDENTIALPROPERTY VALUES: AN ANALYSIS OF THREE SLOW-GROWTH AND THREE FAST-GROWTH CENTRAL CITIES IN THE UNITED STATES

AKRAM M. IJLA

ABSTRACT

Historic designation is thought to have a role in neighborhood economic and community development. Local designation of historic districts is increasingly used as a tool to revitalize deteriorated neighborhoods and to protect endangered historical districts. A number of limitations in several previous studies have made policy development as well as a complete assessment of the impact of designation difficult. Some past studies focused only on historic neighborhoods in one city or one state; other studies have tested the impact of historic designation in general without distinguishing between local, state, or federal designation. Lastly, several earlier studies have also relied on comparing changes in property values in historic areas with those non-historic areas but with too few control variables to isolate the effects of historic area designation. This dissertation expands upon previous work by examining the effects of local historic designation on residential property values across six central cities in five states in the United States while controlling for numerous other variables that could impact the property values. The study employs hedonic regression models and difference on difference (case-control) descriptive statistical models to estimate the impact of local government designation of an area as a historical district on the prices of residential property. This is accomplished by the pairing of each historic district with a similar community that was not designated as historic. The research was performed in three fast-growth and three slow-growth central cities. The results indicate that local historic designation is associated with higher property values in the six central cities. In addition, the positive appreciation effects of local historic designation in slow-growth central cities were higher than in fast-growth central cities by 7.7 percent suggesting that historic designation has a role to play in urban revitalization for areas striving to improve property values despite slow population growth.

Committee:

Dr. Mark Rosentraub (Advisor)

Subjects:

Urban Planning

Keywords:

Historic Preservation; Property Values; Local Historic Designation; Sales Price; Hedonic Regression Model; Positive Axternalities

Whetsel-Ribeau, PaulaRetention of Faculty of Color as it Relates to Their Perceptions of the Academic Climate at Four-Year Predominantly White Public Universities in Ohio
Doctor of Education (Ed.D.), Bowling Green State University, 2007, Leadership Studies
The purpose of this correlational study was to examine the relationships between demographic characteristics, academic climate perceptions, and retention plans of 103 tenured and tenure-track faculty of color at 11 four-year predominantly White public universities in Ohio. The 59-item Faculty Retention Questionnaire was administered online and assessed perceptions of the academic climate defined by six variables (job satisfaction, social climate, faculty-student relationships, role conflict, role clarity, and retention). Demographic characteristics were also measured (e.g., racial/ethnic background, gender, age, sexual orientation, country of origin, institution type, academic discipline, marital status, with/without children, and tenure status). Likert-type scales, multiple choice, and open-ended questions measured employment values and intent to stay in current position. Of the 725 surveys distributed, 103 were submitted, yielding an overall response rate of 14%. Critical Race Theory (CRT) framed this study. Correlational results indicated that job satisfaction was significantly related to and highly important to the retention variable. Analysis of variance revealed that U. S. born faculty of color are more likely to be retained than non-U. S. born. Forward multiple regression analysis identified job satisfaction as the sole predictor of retention with job satisfaction only accounting for 23% of variance in retention. Further regression analysis identified social climate, role clarity, and role conflict as factors that best predict job satisfaction. Conclusions from the study raised larger questions regarding job satisfaction: (1) Does job satisfaction mean something different to faculty of color than it does to mainstream faculty? (2) Do faculty of color perceive job satisfaction as part of their social/cultural experience? (3) Is job satisfaction a part of the dual reality that is inherent in people of color through the identification of being a member of an underrepresented group or by having minority status in America? Responses to these larger questions may be best understood through the recognition and understanding of Critical Race Theory. Findings suggest the importance of providing opportunities for the sharing of subjective cultural worldviews of faculty of color with mainstream faculty with the intent of creating greater understanding, cooperation, and positive relationships, thus serve as a retention strategy. This may provide the opportunity to build an academic climate that supports all faculty. The researcher offers other explanations and suggestions regarding the findings from this study that may be valuable in faculty of color retention.

Committee:

Rachel Vannatta (Advisor)

Keywords:

Critical Race Theory; faculty of color; retention; correlational; multiple regression; diversity; academic climate; higher education; job satisfaction; role conflict; role clarity; role ambiguity; social climate; faculty-student relationships

Chun, YongwanBehavioral specifications of network autocorrelation in migration modeling: an analysis of migration flows by spatial filtering
Doctor of Philosophy, The Ohio State University, 2007, Geography
This research is concerned with the fact that migration flows between two regions are most likely related to other migration flows in a regional system. This phenomenon is called network autocorrelation. Because the presence of network autocorrelation violates the independence assumption that is frequently applied in migration estimation procedures, statistical results without accounting for network autocorrelation are likely to be biased and can potentially produce misleading conclusions. This research aims at [a] investigating the underlying behavioral and structural mechanisms leading to network autocorrelation and its operationalization in a network link matrix, [b] the statistical identification of network autocorrelation in empirical migration systems, [c] the explicit incorporation of network autocorrelation into a migration model by adopting the novel spatial eigenvector filtering approach, and [d] demonstrating the usefulness of the proposed methodology by applying it to interstate migration system of the U.S. during from 1995 to 2000. Network autocorrelation among migration flows can be explained by reflecting on how potential migrants may search for a destination within a regional system. Specifically, as migration is a spatial choice process, it is important to comprehend how migrants perceive space and choose a destination in the space. Network autocorrelation can be incorporated into modeling spatial search behavior of migrants by specifying a proper network dependency structure. In this research, concentrating on competing destination effects and intervening opportunities, two different criteria were proposed: the connectivity through a joint node and the spatial association between nodes. This research proposes spatial eigenvector filtering as a method to model network autocorrelation in Poisson regression. As the spatial eigenvector filtering method is conceptually easy to comprehend and produces robust spatial autocorrelation models, the method can be utilized to isolate network autocorrelation and to control for the effects of network autocorrelation onto a Poisson regression model. As a result, the potential biases in the estimated model parameters and their standard errors are adjusted in the spatially filtered Poisson regression model. The spatial eigenvector filtering approach for network autocorrelation in a migration model is demonstrated by an empirical analysis of the interstate migration in the U.S. during 1995-2000.

Committee:

Morton O'Kelly (Advisor)

Subjects:

Geography

Keywords:

Network autocorrelation; spatial eigenvector filtering; migration; Poisson regression; network link matrix

Geise, Mary JoA Longitudinal Analysis of Outcomes Associated with Ohio's Postsecondary Enrollment Options Program
Doctor of Philosophy (Ph.D.), Bowling Green State University, 2011, Higher Education Administration

Dual enrollment programs, once created for the most advanced students, are now seen as a way to provide an accessible and affordable bridge to postsecondary education for a broader range of students. Research on the outcomes of such programs has been limited in scope and exists for only a few states. This quantitative study analyzed 10 years of postsecondary data from the Ohio Board of Regents to assess outcomes of traditional-aged college students enrolled in the state university system who previously participated in Ohio’s Postsecondary Options Program (PSEOP) as a high school student compared with students of similar academic ability who did not participate in PSEOP. Astin’s I-E-O Model served as the conceptual framework for this study. Several quantitative statistical methods including chi-squared, t-tests, hierarchical logistic regression, and analysis of covariance were used to assess student outcomes.

Ten research questions guided this study, eight of which were successfully answered. The first question descriptively compared demographic and environmental characteristics of students who participated in PSEOP with students who did not participate. The remaining questions investigated significant differences in students’ major field of study choice, first-year retention rates, first-year cumulative grade point average, graduation cumulative grade point average, graduation rates, time-to-degree, and the pursuit of graduate or professional studies within one year of baccalaureate degree attainment. Questions relating to the choice of undergraduate institution and the pursuit of a second major were not answered due to insufficient data to adequately research the outcomes of the two student groups.

Key findings centered on attributes which were significantly related to PSEOP participation and outcomes to which PSEOP participation was a significant contributor. Gender, ethnicity, academic performance, and family characteristics were all related to the decision of whether or not to participate in PSEOP. Students that did participate in PSEOP showed this experience as a significant factor in choosing certain majors and had a statistically significantly shorter time-to-degree completion than those students who did not participate in PSEOP. Results from this study showed areas where participation in PSEOP could be improved, thus widening the access of higher education to a larger pool of students.

Committee:

William E. Knight, PhD (Committee Chair); Kenneth J. Ryan, PhD (Committee Member); Michael D. Coomes, EdD (Committee Member); Robert DeBard, EdD (Committee Member)

Subjects:

Education; Higher Education

Keywords:

Dual enrollment; concurrent enrollment; access to higher education; higher education; accelerated learning options; transitioning to college; public policy; Ohio higher education; Ohio K-12 education; hierarchical logistic regression

Kim, YoungkookImpacts of Transportation, Land Uses, and Meteorology on Urban Air Quality
Doctor of Philosophy, The Ohio State University, 2010, City and Regional Planning

Criteria air pollutants, such as nitrogen oxides (NOx), carbon monoxide (CO), sulfur dioxide (SO2), particulate matter (PM), and ozone (O3), are characterized by temporal and locational hot spots in urban areas, frequently violating pollution standards, and, as a result, threatening the health and well-being of the population. Several factors, such as the intensity and duration of emissions, the chemical reactions among pollutants, the uptake and assimilation of pollutants by urban vegetation, and the meteorological factors that induce chemical reactions and atmospheric dispersion, have been considered as explanatory variables in air quality models. Among them, emissions from motor vehicles turn out to be a key determinant of the spatial and temporal patterns of ambient pollution concentrations.

The purpose of this research is to formulate and estimate (1) metropolitan-wide time-series air quality models and (2) land-use regression (LUR) air quality panel models, in order to explain spatio-temporal variations in pollution concentrations. Using the Seoul Metropolitan Area as a case study, traffic counts, vehicle-kilometers-traveled (VKT), land uses, and meteorological factors, such as solar radiation, temperature, humidity, wind speed and wind direction, are used as explanatory variables. An extensive understanding of atmospheric pollutants chemistry is reflected in the formulation of these models. Differences in concentrations measured at air quality monitoring stations (AQMs) across the week (weekdays vs. Sunday) and geographical locations (roadside vs. background), are also investigated using dummy variables and the product of these variables with the original variables.

The results of the time-series models and panel regression models indicate that traffic counts and VKT are significant in explaining the concentrations of both directly emitted pollutants, such as NO2, CO, SO2, and PM, and O3, a secondary pollutant. The concentrations of the directly emitted pollutants are positively correlated with vehicle flows. In the case of O3, however, vehicle emissions have a negative impact on O3 concentrations. Since solar radiation, temperature, humidity, and wind speed influence both chemical reactions and physical dispersion, these factors are closely correlated with pollution concentrations. In particular, solar radiation plays a critical role regarding NO2 and O3 concentrations. Ultraviolet (UV) radiation causes the photodissociation of NO2, providing favorable conditions for the generation of O3 in the troposphere. The estimated models confirm that solar radiation have a positive effect in the O3 models, and a negative one in the NO2 models.

Reduced traffic flows on Sunday increase the ratio of volatile organic compounds (VOC) to NOx and, consequently, lead to favorable conditions for O3 generation. Less O3 titration and less HNO3 formation occur on Sunday as compared to weekdays, resulting in higher O3 concentrations on Sunday. For directly emitted pollutants, reduced traffic flows on Sunday induce a decrease in concentrations. The locations of the AQMs turn out to be critical. Traffic flows closer to AQMs have larger impacts on pollution concentrations. The product terms between VKT and roadside dummy variable display the expected results: for directly emitted pollutants, the coefficients are significant and positive, suggesting that the impacts of roadside VKT are greater than those of background VKT. In the case of O3, the estimated coefficients are negative, indicating that the negative impacts of VKT on O3 concentrations are increased at roadside areas. Nitric oxide emissions from commercial and residential areas have a negative impact on O3 concentrations. Plants have an O3 assimilation capacity, but also emit biogenic VOC during the growing seasons, generating simultaneous negative and positive impacts. The overall vegetative areas impact on O3 concentrations is positive. For directly emitted pollutants, however, vegetative areas have a negative impact. Since residential, commercial, and industrial areas generate anthropogenic emissions, the coefficients of these land uses are positive.

Committee:

Jean-Michel Guldmann, PhD (Committee Chair); Steven Gordon, PhD (Committee Member); Philip Viton, PhD (Committee Member); Gulsah Akar, PhD (Committee Member)

Subjects:

Urban Planning

Keywords:

air pollution; time-series model; panel model; land-use regression

Dougherty, Andrew W.Intelligent Design of Metal Oxide Gas Sensor Arrays Using Reciprocal Kernel Support Vector Regression
Doctor of Philosophy, The Ohio State University, 2010, Physics

Metal oxides are a staple of the sensor industry. The combination of their sensitivity to a number of gases, and the electrical nature of their sensing mechanism, make the particularly attractive in solid state devices. The high temperature stability of the ceramic material also make them ideal for detecting combustion byproducts where exhaust temperatures can be high. However, problems do exist with metal oxide sensors. They are not very selective as they all tend to be sensitive to a number of reduction and oxidation reactions on the oxide’s surface. This makes sensors with large numbers of sensors interesting to study as a method for introducing orthogonality to the system. Also, the sensors tend to suffer from long term drift for a number of reasons.

In this thesis I will develop a system for intelligently modeling metal oxide sensors and determining their suitability for use in large arrays designed to analyze exhaust gas streams. It will introduce prior knowledge of the metal oxide sensors’ response mechanisms in order to produce a response function for each sensor from sparse training data. The system will use the same technique to model and remove any long term drift from the sensor response. It will also provide an efficient means for determining the orthogonality of the sensor to determine whether they are useful in gas sensing arrays.

The system is based on least squares support vector regression using the reciprocal kernel. The reciprocal kernel is introduced along with a method of optimizing the free parameters of the reciprocal kernel support vector machine. The reciprocal kernel is shown to be simpler and to perform better than an earlier kernel, the modified reciprocal kernel. Least squares support vector regression is chosen as it uses all of the training points and an emphasis was placed throughout this research for extracting the maximum information from very sparse data.

The reciprocal kernel is shown to be effective in modeling the sensor responses in the time, gas and temperature domains, and the dual representation of the support vector regression solution is shown to provide insight into the sensor’s sensitivity and potential orthogonality. Finally, the dual weights of the support vector regression solution to the sensor’s response are suggested as a fitness function for a genetic algorithm, or some other method for efficiently searching large parameter spaces.

Committee:

Bruce Patton, PhD (Advisor); Ralf Bundschuh, PhD (Committee Member); David Stroud, PhD (Committee Member); Patricia Morris, PhD (Committee Member); Edward Overman, PhD (Committee Member)

Subjects:

Artificial Intelligence; Materials Science; Physics

Keywords:

support vector regression; metal oxide; sensor arrays

Kalapati, Raga SAnalysis of Ozone Data Trends as an Effect of Meteorology and Development of Forecasting Models for Predicting Hourly Ozone Concentrations and Exceedances for Dayton, OH, Using MM5 Real-Time Forecasts
Master of Science, University of Toledo, 2004, Civil Engineering
The objective of this research was to develop and evaluate models for predicting hourly ozone concentrations, ozone exceedances and hourly air quality index (AQI) in Dayton, OH. As the hourly ozone concentrations are closely related to the meteorological conditions, three variables- temperature, wind speed, and dew point temperature- were chosen for this study. The ozone data were extracted from the EPA’s AIRS database for the period 1996-2003. The meteorological data was taken from the National Climatic Data Center (NCDC) for the same period. An analysis of variations in hourly ozone concentrations and ozone episode occurrences was carried out for the period Apr.-Oct. for the years 1996-1999. Also, analysis of the long-term trends in annual means of ozone concentrations, temperature, wind speed, and dew point temperature was performed using the same data set. Based on this analysis, the ozone data was divided into pre-summer (Apr.-Jul.) and post-summer (Aug.-Oct.) seasons, to account for seasonal variations, and each season was further divided into three regimes, namely, stable period (hours: 1-8), ascent period (hours: 9-16), and descent period (hours: 17-24). The KZ filter technique was used to reduce the scatter in the time series, and models were developed for the three regimes for each season by regression, using the corresponding independent parameter values. A total of twelve models were developed to predict ozone concentrations for pre-summer and post-summer periods. Six models considered temperature, wind speed, and dew point temperature as the independent variables (three-parameter models), and the other six considered temperature and wind speed as variables (two-parameter models). Also, three models each for pre-summer and post-summer season were developed for predicting the ozone exceedances. The performance of the models was evaluated in three ways: a) Initial evaluation (or validation) of the models was conducted using 2002 data. b) The effectiveness of these models was further evaluated using available MM5 (a mesoscale meteorological forecasting model) real-time forecasts from the Ohio State University for the months of Aug.-Oct., 2003. c) Finally, the performance of the three-parameter models was compared with that of the two-parameter models. All the evaluations were made using statistical evaluation parameters discussed later. The study shows that the forecasts of hourly ozone concentrations made by the models based on KZ filters are reliable only to a limited extent. However, the models performed well in predicting AQI values reported by the EPA. Also, the three-parameter models performed better in predicting the peak concentrations when compared to the two-parameter models.

Committee:

Ashok Kumar (Advisor)

Subjects:

Engineering, Environmental

Keywords:

hourly ozone concentrations;ozone exceedances;hourly air quality index;Ozone annual and seasonal trends;MM5 (a mesoscale meteorological forecasting model);KZ filter technique;regression analysis;statistical evaluation parameters

GREEN, CHRISTOPHER FRANKASSESSMENT AND MODELING OF INDOOR AIR QUALITY
PhD, University of Cincinnati, 2002, Engineering : Environmental Science
Exposure to contaminated air has become an increased problem due to a variety of factors. Stachybotrys chartarum and other potentially harmful microorganisms implicated with indoor air problems have garnered national attention in recent years. Accurate assessment of indoor contaminated air, however, has proven to be prohibitively labor, time, cost, and training intensive. This dissertation presents the development of a model that accurately predicts the levels of indoor air biological contaminants using a number of independent variables that can be quickly calculated without expensive, time-consuming scientific techniques. Thirty-nine (39) residences were sampled in the Greater Cincinnati area using Andersen 2-stage air samplers loaded with Malt Extract Agar, Trypicase Soy Agar, Czapek’s Cellulose Agar, and Corn Meal Agar. After air sampling, the Petri dishes were incubated, the number of colonies from each plate were enumerated and the total number of viable cells/m3 were calculated. Initial walk-through of each site included an investigation of any possible presence of fungi or evidence of water damage. Relative humidity and temperature inside the site and out were recorded. Building site, size, and type were also noted and residents of the house were given an indoor health quality questionnaire to fill out. Independent variables were then compared together to the dependent variable using multiple linear regression. This was done using Analyze-it <@reg> for Microsoft Excel <@reg>. The model compared independent variables like temperature and humidity to fungal and bacterial bioaerosol counts obtained from the Andersen samplers. The final air model predicts indoor cell counts with 96% accuracy. The goal for this model was 90% accuracy.

Committee:

Dr. Pasquale Scarpino (Advisor)

Keywords:

indoor air quality; bioaerosols; fungal indentification; mmultiple linear regression; modeling

Jennings, Alan LanceAutonomous Motion Learning for Near Optimal Control
Doctor of Philosophy (Ph.D.), University of Dayton, 2012, Electrical Engineering

Human intelligence has appealed to the robotics community for a long time; specifically, a person's ability to learn new tasks efficiently and eventually master the task. This ability is the result of decades of development as a person matures from an infant to an adult and a similar developmental period seems to be required if robots are to obtain the ability to learn and master new skills. Applying developmental stages to robotics is a field of study that has been growing in acceptance. The paradigm shift is from directly pursuing the desired task to progressively building competencies until the desired task is reached. This dissertation seeks to apply a developmental approach to autonomous optimization of robotic motions, and the methods presented extend to function shaping and parameter optimization.

Humans have a limited ability to concentrate on multiple tasks at once. For robots with many degrees of freedom, human operators need a high-level interface, rather than controlling the positions of each angle. Motion primitives are scalable control signals that have repeatable, high-level results. Examples include walking, jumping or throwing where the result can be scaled in terms of speed, height or distance. Traditionally, motion primitives require extensive, robot-specific analysis making development of large databases of primitives infeasible. This dissertation presents methods of autonomously creating and refining optimal inverse functions for use as motion primitives. By clustering contiguous local optima, a continuous inverse function can be created by interpolating results. The additional clusters serve as alternatives if the chosen cluster is poorly suited to the situation. For multimodal problems, a population based optimization can efficiently search a large space.

Staged learning offers a path to mimic the progression from novice to master, as seen in human learning. The dimension of the input wave parameterization, which is the number degrees of freedom for optimization, is incremented to allow for additional improvement. As the parameterization increases in order, the true optimal continuous-time control signal is approached. All previous experiences can be directly moved to the higher parameterization when expanding the parameterization, if a proper parameterization is selected. Incrementally increasing complexity and retaining experience efficiently optimizes to high dimensions when contrasted with undirected global optimizations, which would need to search the entire high dimension space. The method presented allows for unbounded resolution since the parameterization is not fixed at programming.

This dissertation presents several methods that make steps towards the goal of learning and mastering motion-related tasks without programmed, task-specific heuristics. Trajectory optimization based on a high-level system description has been demonstrated for a robotic arm performing a pick-place task. In addition, the inverse optimal function was applied to optimizing robotic tracking precision in a method suitable for online tracking. Staging of the learning is able determine an optimal motor spin-up waveform despite large variations in system parameters. Global optimization, using a population based search, and unbounded resolution increasing provide the foundation for autonomously developing scalable motions superior to what can be designed by hand.

Committee:

Raúl Ordóñez, Ph. D. (Advisor); Frederick G. Harmon, Ph. D., Lt Col (Committee Member); Eric Balster, Ph. D. (Committee Member); Andrew Murray, Ph. D. (Committee Member)

Subjects:

Applied Mathematics; Artificial Intelligence; Electrical Engineering; Robotics; Robots

Keywords:

Nonlinear Optimization; Optimal Control; Developmental Learning; Robotics; Inverse Functions; Locally Weighted Regression

Fulkerson, Matthew DGas Sensor Array Modeling and Cuprate Superconductivity From Correlated Spin Disorder
Doctor of Philosophy, The Ohio State University, 2002, Physics

In part I, a kernel regression method is developed for modeling gas sensor response functions, for the purpose of identifying the composition of gas mixtures. A quantitative measure of orthogonality in sensor arrays is introduced. The method is applied to TiO2 sensor arrays exposed to O2 and CO. A sensing mechanism is analyzed by applying Wolkenstein's theory of chemisorption to the single grain TiO2 response to O2 and CO. It is found that this mechanism, in combination with Wolkenstein's theory, can adequately describe the response of granular TiO2 sensors.

In part II, a spin-dependent tight-binding model for the hole-doped cuprates is developed. The model incorporates the effects of antiferromagnetism and spin disorder. In the coherent potential approximation, the effective medium Green's function is derived and single-particle properties are obtained. When the spin disorder is correlated, an effective interaction between holes leads to dx2-y2 superconductivity. The pseudogap region of the cuprate phase diagram is interpreted within this theory.

Committee:

Bruce Patton (Advisor)

Subjects:

Physics, Condensed Matter

Keywords:

TiO2; sensor arrays; kernel regression; coherent potential; spin disorder; cuprate superconductivity

Next Page