Physical Sciences and Mathematics Commons^™

Simulations Of A New Response-Adaptive Biased Coin Design, Aleksandra Stein

Department of Statistics: Dissertations, Theses, and Student Work

Modern medical experiments accrue and treat patients--hence obtain treatment response data--throughout a trial. Designs which prospectively plan to modify patient allocation by leveraging accumulating data are response-adaptive randomization (RAR) designs. Many such designs attempt to balance the desire to bias assignment proportions towards a treatment which is performing better against the need to maintain randomization in the face of continued equipoise.

This dissertation consists of simulated investigations into frequentist and ethical properties of an new RAR biased coin design. Chapter 2 proposes a new adaptive design for phase III clinical trials, a modification of the 2001 Bandyopadhyay and Biswas biased …

Beta-Binomial Kriging: A New Approach To Modeling Spatially Correlated Proportions, Aimee Schwab

Department of Statistics: Dissertations, Theses, and Student Work

Spatially correlated count data sets appear often in applied data analysis problems, but there is little consensus in the literature about how best to analyze the data. The two prevailing approaches provide accurate parameter estimates and predictions, at the cost of model interpretability and simplicity. This dissertation will present a new approach to modeling spatially correlated binomial observations: beta-binomial kriging. The model proposed here is a modified form of spatial kriging which assumes the data are generated from a correlated beta-binomial distribution. Given this assumption, the spatial parameters and predicted values can be estimated using simple matrix algebra. Beta-binomial kriging …

A New Approach To Modeling Multivariate Time Series On Multiple Temporal Scales, Tucker Zeleny

Department of Statistics: Dissertations, Theses, and Student Work

In certain situations, observations are collected on a multivariate time series at a certain temporal scale. However, there may also exist underlying time series behavior on a larger temporal scale that is of interest. Often times, identifying the behavior of the data over the course of the larger scale is the key objective. Because this large scale trend is not being directly observed, describing the trends of the data on this scale can be more difficult. To further complicate matters, the observed data on the smaller time scale may be unevenly spaced from one larger scale time point to the …

A Comparison Of Population-Averaged And Cluster-Specific Approaches In The Context Of Unequal Probabilities Of Selection, Natalie A. Koziol

College of Education and Human Sciences: Dissertations, Theses, and Student Research

Sampling designs of large-scale, federally funded studies are typically complex, involving multiple design features (e.g., clustering, unequal probabilities of selection). Researchers must account for these features in order to obtain unbiased point estimators and make valid inferences about population parameters. Single-level (i.e., population-averaged) and multilevel (i.e., cluster-specific) methods provide two alternatives for modeling clustered data. Single-level methods rely on the use of adjusted variance estimators to account for dependency due to clustering, whereas multilevel methods incorporate the dependency into the specification of the model.

Although the literature comparing single-level and multilevel approaches is vast, comparisons have been limited to the …

Global Network Inference From Ego Network Samples: Testing A Simulation Approach, Jeffrey A. Smith

Department of Sociology: Faculty Publications

Network sampling poses a radical idea: that it is possible to measure global network structure without the full population coverage assumed in most network studies. Network sampling is only useful, however, if a researcher can produce accurate global network estimates. This article explores the practicality of making network inference, focusing on the approach introduced in Smith (2012). The method uses sampled ego network data and simulation techniques to make inference about the global features of the true, unknown network. The validity check here includes more difficult scenarios than previous tests, including those that go beyond the initial scope conditions of …

Best Practice Recommendations For Data Screening, Justin A. Desimone, Peter D. Harms, Alice J. Desimone

Department of Management: Faculty Publications

Survey respondents differ in their levels of attention and effort when responding to items. There are a number of methods researchers may use to identify respondents who fail to exert sufficient effort in order to increase the rigor of analysis and enhance the trustworthiness of study results. Screening techniques are organized into three general categories, which differ in impact on survey design and potential respondent awareness. Assumptions and considerations regarding appropriate use of screening techniques are discussed along with descriptions of each technique. The utility of each screening technique is a function of survey design and administration. Each technique has …

Genomic-Enabled Prediction Of Ordinal Data With Bayesian Logistic Ordinal Regression, Osval A. Montesinos-López, Abelardo Montesinos-López, José Crossa, Juan Burgueño, Kent M. Eskridge

Department of Statistics: Faculty Publications

Most genomic-enabled prediction models developed so far assume that the response variable is continuous and normally distributed. The exception is the probit model, developed for ordered categorical phenotypes. In statistical applications, because of the easy implementation of the Bayesian probit ordinal regression (BPOR) model, Bayesian logistic ordinal regression (BLOR) is implemented rarely in the context of genomic-enabled prediction [sample size (n) is much smaller than the number of parameters (p)]. For this reason, in this paper we propose a BLOR model using the Pólya-Gamma data augmentation approach that produces a Gibbs sampler with similar full conditional distributions of the BPORmodel …

Establishment And Persistence Of Yellow-Flowered Alfalfa No-Till Interseeded Into Crested Wheatgrass Stands, Christopher G. Misar, Lan Xu, Roger N. Gates, Arvid Boe, Patricia S. Johnson, Christopher S. Schauer, John R. Rickertsen, Walter Stroup

Department of Statistics: Faculty Publications

Crested wheatgrass [Agropyron cristatum (L.) Gaertn., A. desertorum

(Fisch. ex Link) Schult., and related taxa] often exists

in near monoculture stands in the northern Great Plains.

Introducing locally adapted yellow-flowered alfalfa [Medicago

sativa L. subsp. falcata (L.) Arcang.] would complement crested

wheatgrass. Our objective was to evaluate effects of seeding

date, clethodim {(E) -2-[1-[[(3-chloro-2-propenyl)oxy]imino]

propyl]-5-[2-(ethylthio)propyl]-3-hydroxy-2-cyclohexen-1-one}

sod suppression, and seeding rate on initial establishment and

stand persistence of Falcata, a predominantly yellow-flowered

alfalfa, no-till interseeded into crested wheatgrass. Research was

initiated in August 2008 at Newcastle, WY; Hettinger, ND;

Fruitdale, SD; and Buffalo, SD. Effects of treatment …

Effect Of Dexamethasone Prodrug On Inflamed Temporomandibular Joints In Juvenile Rats, Mitchell Knudsen, Matthew Bury, Callie Holwegner, Adam L. Reinhardt, Fang Yuan, Yijia Zhang, Peter Giannini, D. B. Marx, Dong Wang, Richard A. Reinhardt

Department of Statistics: Faculty Publications

Introduction: Juvenile idiopathic arthritis (JIA) often causes inflammation of the temporomandibular joint (TMJ) and has been treated with both systemic and intra-articular steroids, with concerns about effects on growing bones. In this study, we evaluated the impact of a macromolecular prodrug of dexamethasone (P-DEX) with inflammation-targeting potential applied systemically or directly to the TMJ.

Methods: Joint inflammation was initiated by injecting two doses of complete Freund’s adjuvant (CFA) at 1-month intervals into the right TMJs of 24 growing Sprague–Dawley male rats (controls on left side). Four additional rats were not manipulated. With the second CFA injection, animals received (1) 5 …

Threshold Models For Genome-Enabled Prediction Of Ordinal Categorical Traits In Plant Breeding, Osval A. Montesinos-López, Abelardo Montesinos-López, Paulino Pérez-Rodríguez, Gustavo De Los Campos, Kent M. Eskridge, José Crossa

Department of Statistics: Faculty Publications

Categorical scores for disease susceptibility or resistance often are recorded in plant breeding. The aim of this study was to introduce genomic models for analyzing ordinal characters and to assess the predictive ability of genomic predictions for ordered categorical phenotypes using a threshold model counterpart of the Genomic Best Linear Unbiased Predictor (i.e., TGBLUP). The threshold model was used to relate a hypothetical underlying scale to the outward categorical response. We present an empirical application where a total of nine models, five without interaction and four with genomic x environment interaction (G·E) and genomic additive x additive x environment interaction …

Measuring Peer Socialization For Adolescent Substance Use:A Comparison Of Perceived And Actual Friends’ Substance Use Effects, Arielle R. Deutsch, Pavel Chernyavskiy, Douglas Steinley, Wendy S. Slutske

Department of Statistics: Faculty Publications

Objective: There has been an increase in the use of social network analysis in studies of peer socialization effects on adolescent substance use. Some researchers argue that social network analyses provide more accurate measures of peer substance use, that the alternate strategy of assessing perceptions of friends’ drug use is biased, and that perceptions of peer use and actual peer use represent different constructs. However, there has been little research directly comparing the two effects, and little is known about the extent to which the measures differ in the magnitude of their influence on adolescent substance use, as well as …

A Copula Based Approach For Design Of Multivariate Random Forests For Drug Sensitivity Prediction, Saad Haider, Raziur Rahman, Souparno Ghosh, Ranadip Pal

Department of Statistics: Faculty Publications

Modeling sensitivity to drugs based on genetic characterizations is a significant challenge in the area of systems medicine. Ensemble based approaches such as Random Forests have been shown to perform well in both individual sensitivity prediction studies and team science based prediction challenges. However, Random Forests generate a deterministic predictive model for each drug based on the genetic characterization of the cell lines and ignores the relationship between different drug sensitivities during model generation. This application motivates the need for generation of multivariate ensemble learning techniques that can increase prediction accuracy and improve variable importance ranking by incorporating the relationships …

Design Of Probabilistic Random Forests With Applications To Anticancer Drug Sensitivity Prediction, Raziur Rahman, Saad Haider, Souparno Gosh, Ranadip Pal

Department of Statistics: Faculty Publications

Random forests consisting of an ensemble of regression trees with equal weights are frequently used for design of predictive models. In this article, we consider an extension of the methodology by representing the regression trees in the form of probabilistic trees and analyzing the nature of heteroscedasticity. The probabilistic tree representation allows for analytical computation of confidence intervals (CIs), and the tree weight optimization is expected to provide stricter CIs with comparable performance in mean error. We approached the ensemble of probabilistic trees’ prediction from the perspectives of a mixture distribution and as a weighted sum of correlated random variables. …

S1: Supplementary Information For Article: A Copula Based Approach For Design Of Multivariate Random Forests For Drug Sensitivity Prediction, Saad Haider, Raziur Rahman, Souparno Ghosh, Ranadip Pal

Department of Statistics: Faculty Publications

Changes in performance with prior feature selection

Random forest (RF) is designed to create uncorrelated trees using random subsets of features in each node of each tree. RF by itself is a great tool for feature selection from a high dimensional set of features. But we observed that the prediction accuracy is improved when a prior feature selection (RELIEFF) [1] approach is implemented. Table A shows the performance of RF, VMRF and CMRF with and without RELIEFF feature selection in 2 drug sets of GDSC.

Performance Analysis for drugsets consisting of more 8 than two drugs

We have generated empirical …

Voc Emissions From Beef Feedlot Pen Surfaces As Affected By Within-Pen Location, Moisture And Temperature, Bryan L. Woodbury, John E. Gilley, David B. Parker, David B. Marx, Roger A. Eigenberg

Department of Statistics: Faculty Publications

A laboratory study was conducted to evaluate the effects of pen location, moisture, and temperature on emissions of volatile organic compounds (VOC) from surface materials obtained from feedlot pens where beef cattle were fed a diet containing 30% wet distillers grain plus solubles. Surface materials were collected from the feed trough (bunk), drainage, and raised areas (mounds) within three feedlot pens. The surface materials were mixed with water to represent dry, wet, or saturated conditions and then incubated at temperatures of 5, 15, 25 and 35 C. A wind tunnel and gas chromatograph-mass spectrometer were used to collect and quantify …

Spin Glass Reflection Of The Decoding Transition For Quantum Error Correcting Codes, Alexey Kovalev, Leonid P. Pryadko

Department of Physics and Astronomy: Faculty Publications

We study the decoding transition for quantum error correcting codes with the help of a mapping to random-bond Wegner spin models. Families of quantum low density parity-check (LDPC) codes with a finite decoding threshold lead to both known models (e.g., random bond Ising and random plaquette Z₂ gauge models) as well as unexplored earlier generally non-local disordered spin models with non-trivial phase diagrams. The decoding transition corresponds to a transition from the ordered phase by proliferation of "post-topological" extended defects which generalize the notion of domain walls to non-local spin models. In recently discovered quantum LDPC code families with …

Modeling The Dynamic Processes Of Challenge And Recovery (Stress And Strain) Over Time, Fan Yang

Department of Statistics: Dissertations, Theses, and Student Work

A dynamic process with challenge and recovery is an important branch in the family of stochastic processes. The dependent data of such processes are often observed over time, and hence, are time dependent. The purpose of this dissertation is to develop methods to characterize a dynamic process with challenge and recovery under different dimensionalities and error assumptions. In this dissertation, a univariate dynamic process under Gaussian assumption is discussed first and a bi-logistic model is developed by three different methods: compartment, additive, and Bayesian. Then the discussion is extended to a bivariate hysteresis system with challenge and recovery. Three methods: …

Equate: Observed-Score Linking And Equating In R, Anthony D. Albano

Department of Educational Psychology: Faculty Publications

Linking and equating are statistical procedures used to convert scores from one measurement scale to another. These procedures are most often used in testing programs that involve multiple test forms, where adjustments are made for form difficulty differences when creating a measurement scale that is common across forms. Linking and equating methods are traditionally distinguished by the type of scores they are applied to, whether observed scores or scores from an item response theory model. Methods are also distinguished by the study design under which measurements are taken. The R package equate (Albano, 2014) is free, open-source software for conducting …