Statistics and Probability Commons^™

Visualizing Longitudinal Data With Dropouts, Mithat Gonen

Memorial Sloan-Kettering Cancer Center, Dept. of Epidemiology & Biostatistics Working Paper Series

A triangle plot is proposed to display longitudinal data with dropouts. The triangle plot is a tool of data visualization that can also serve as a graphical check for informativeness of the dropout process. There are similarities between the lasagna plot and the triangle plot but the explicit use of dropout time as an axis is an advantage of the triangle plot over the more commonly used graphical strategies for longitudinal data. It is possible to interpret the triangle plot as a trellis plot 1 which gives rise to several extensions such as the triangle histogram and the triangle boxplot. …

On The Simulation Of Longitudinal Discrete Data With Specified Marginal Means And First-Order Antedependence, Matthew Guerra, Justine Shults

UPenn Biostatistics Working Papers

We propose a straightforward approach for simulation of discrete random variables with overdispersion, specified marginal means, and product correlations that are plausible for longitudinal data with equal, or unequal, temporal spacings. The method stems from results we prove for variables with first-order antedependence and linearity of the conditional expectations. The proposed approach will be especially useful for assessment of methods such as generalized estimating equations, which specify separate models for the marginal means and correlation structure of measurements on a subject.

Mixtures Of Receiver Operating Characteristic Curves, Mithat Gonen

Memorial Sloan-Kettering Cancer Center, Dept. of Epidemiology & Biostatistics Working Paper Series

Rationale and Objectives: ROC curves are ubiquitous in the analysis of imaging metrics as markers of both diagnosis and prognosis. While empirical estimation of ROC curves remains the most popular method, there are several reasons to consider smooth estimates based on a parametric model.

Materials and Methods: A mixture model is considered for modeling the distribution of the marker in the diseased population motivated by the biological observation that there is more heterogeneity in the diseased population than there is in the normal one. It is shown that this model results in an analytically tractable ROC curve which is itself …

The Effects Of Variable Viscosity On The Peristaltic Flow Of Non-Newtonian Fluid Through A Porous Medium In An Inclined Channel With Slip Boundary Conditions, Ambreen Afsar Khan, R. Ellahi, Muhammad Usman

Mathematics Faculty Publications

The present paper investigates the peristaltic motion of an incompressible non-Newtonian fluid with variable viscosity through a porous medium in an inclined symmetric channel under the effect of the slip condition. A long wavelength approximation is used in mathematical modeling. The system of the governing nonlinear partial differential equation has been solved by using the regular perturbation method and the analytical solutions for velocity and pressure rise have been obtained in the form of stream function. In the obtained solution expressions, the long wavelength and low Reynolds number assumptions are utilized. The salient features of pumping and trapping phenomena are …

A Superposed Log-Linear Failure Intensity Model For Repairable Artillery Systems, Byeong Min Mun, Suk Joo Bae, Paul Kvam

Department of Math & Statistics Faculty Publications

This article investigates complex repairable artillery systems that include several failure modes. We derive a superposed process based on a mixture of nonhomogeneous Poisson processes in a minimal repair model. This allows for a bathtub-shaped failure intensity that models artillery data better than currently used methods. The method of maximum likelihood is used to estimate model parameters and construct confidence intervals for the cumulative intensity of the superposed process. Finally, we propose an optimal maintenance policy for repairable systems with bathtub-shaped intensity and apply it to the artillery-failure data.

The Beverton–Holt Q-Difference Equation, Martin Bohner, Rotchana Chieochan

Mathematics and Statistics Faculty Research & Creative Works

The Beverton–Holt model is a classical population model which has been considered in the literature for the discrete-time case. Its continuous-time analogue is the well-known logistic model. In this paper, we consider a quantum calculus analogue of the Beverton–Holt equation. We use a recently introduced concept of periodic functions in quantum calculus in order to study the existence of periodic solutions of the Beverton–Holt q-difference equation. Moreover, we present proofs of quantum calculus versions of two so-called Cushing–Henson conjectures. © 2013 The Author(s). Published by Taylor & Francis.

Drosophila Melanogaster Selection For Survival After Infection With Bacillus Cereus Spores: Evolutionary Genetic And Phenotypic Investigations Of Respiration And Movement, Junjie Ma, Andrew K. Benson, Stephen D. Kachman, Deidra J. Jacobsen, Lawrence G. Harshman

Department of Statistics: Faculty Publications

Laboratory populations of D.melanogaster have been subjected to selection for survival after live spores of B. cereus were introduced as a pathogenic agent. The present study was designed to investigate correlated traits: respiration as ametabolic trait and movement as a behavioral trait.An underlying hypothesiswas that the evolution of increased survival after B. cereus infection exerts ametabolic cost associated with elevated immunity and this would be detected by increased respiration rates. There was support for this hypothesis in the male response to selection, but not for selected-line females. Two phenotypic effects were also observed in the study. Females especially showed a …

The Change In Differing Leukocyte Populations During Vaccination To Bovine Respiratory Disease And Their Correlations With Lung Scores, Health Records, And Average Daily Gain, R. J. Leach, C. G. Chitko-Mckown, G. L. Bennett, S. A. Jones, Stephen D. Kachman, J. W. Keele, K. A. Leymaster, R. M. Thallman, L. A. Kuehn

Department of Statistics: Faculty Publications

Bovine respiratory disease (BRD) is the most economically important disease in U.S. feedlots. Infection can result in morbidity, mortality, and reduced average daily gain. Cheap and reliable genetic methods of prediction and protection from BRD would be highly advantageous to the industry. The immune response may correlate with BRD incidence. Cattle (n = 2,182) were vaccinated against common viral and bacterial pathogens of BRD. Two blood samples were collected, one during booster vaccination and one 21d later, enabling 3 phenotypes for each trait [prebooster (pre), postbooster (post), and delta (post minus pre)]. From the blood samples innate and adaptive responses …

Prediction In M-Complete Problems With Limited Sample Size, Jennifer Lynn Clarke, Bertrand Clarke, Chi-Wai Yu

Department of Statistics: Faculty Publications

We define a new Bayesian predictor called the posterior weighted median (PWM) and compare its performance to several other predictors including the Bayes model average under squared error loss, the Barbieri-Berger median model predictor, the stacking predictor, and the model average predictor based on Akaike's information criterion. We argue that PWM generally gives better performance than other predictors over a range of M-complete problems. This range is between the M-closed-M-complete boundary and the M-complete- M-open boundary. Indeed, as a problem gets closer to M-open, it seems that M-complete predictive methods begin to break down. Our comparisons rest on extensive simulations …

Generalized Least-Squares Regressions Ii: Theory And Classification, Nataniel Greene

Publications and Research

In the first paper of this series, a variety of known and new symmetric and weighted least-squares regression methods were presented with efficient derivations. This paper continues and generalizes the previous work with a theory for deriving, analyzing, and classifying all symmetric and weighted least-squares regression methods.

Food For Thought … Mechanistic Validation, Thomas Hartung, Sebastian Hoffman, Martin Stephens

Experimentation Collection

Validation of new approaches in regulatory toxicology is commonly defined as the independent assessment of the reproducibility and relevance (the scientific basis and predictive capacity) of a test for a particular purpose. In large ring trials, the emphasis to date has been mainly on reproducibility and predictive capacity (comparison to the traditional test) with less attention given to the scientific or mechanistic basis. Assessing predictive capacity is difficult for novel approaches (which are based on mechanism), such as pathways of toxicity or the complex networks within the organism (systems toxicology). This is highly relevant for implementing Toxicology for the 21st …

Investigation Of A Pregnancy Lifestyle Intervention Using Mediation Analysis And A Power Analysis Simulation, Kelsey Grantham

Statistics

No abstract provided.

A Frailty Approach For Survival Analysis With Error-Prone Covariate, Sehee Kim, Yi Li, Donna Spiegelman

The University of Michigan Department of Biostatistics Working Paper Series

This paper discovers an inherent relationship between the survival model with covariate measurement error and the frailty model. The discovery motivates our using a frailty-based estimating equation to draw inference for the proportional hazards model with error-prone covariates. Our established framework accommodates general distributional structures for the error-prone covariates, not restricted to a linear additive measurement error model or Gaussian measurement error. When the conditional distribution of the frailty given the surrogate is unknown, it is estimated through a semiparametric copula function. The proposed copula-based approach enables us to fit flexible measurement error models without the curse of dimensionality as …

Ultrahigh Dimensional Time Course Feature Selection, Peirong Xu, Lixing Zhu, Yi Li

The University of Michigan Department of Biostatistics Working Paper Series

Statistical challenges arise from modern biomedical studies that produce time course genomic data with ultrahigh dimensions. In a renal cancer study that motivated this paper, the pharmacokinetic measures of a tumor suppressor (CCI-779) and expression levels of 12625 genes were measured for each of 33 patients at 8 and 16 weeks after the start of treatments, with the goal of identifying predictive gene transcripts and the interactions with time in peripheral blood mononuclear cells for pharmacokinetics over the time course. The resulting dataset defies analysis even with regularized regression. Although some remedies have been proposed for both linear and generalized …

Selection Of Latent Variables For Multiple Mixed-Outcome Models, Ling Zhou, Huazhen Lin, Xin-Yuan Song, Yi Li

The University of Michigan Department of Biostatistics Working Paper Series

Latent variable models have been widely used for modeling the dependence structure of multiple outcomes data. As the formulation of a latent variable model is often unknown a priori, misspecification could distort the dependence structure and lead to unreliable model inference. More- over, the multiple outcomes are often of varying types (e.g., continuous and ordinal), which presents analytical challenges. In this article, we present a class of general latent variable models that can accommodate mixed types of outcomes, and further propose a novel selection approach that simultaneously selects latent variables and estimates model parameters. We show that the proposed estimators …

Semiparametric Latent Variable Transformation Models For Multiple Mixed Outcomes, Huazhen Lin, Ling Zhou, Robert Elashoff, Yi Li

The University of Michigan Department of Biostatistics Working Paper Series

No abstract provided.

Semiparametric Transformation Models For Semicompeting Survival Data, Huazhen Lin, Ling Zhou, Chunhong Li, Yi Li

The University of Michigan Department of Biostatistics Working Paper Series

Semicompeting risk outcome data, e.g. time to disease progression and time to death, are commonly collected in clinical trials, but complicated analytical tools hamper the analysis and the interpretation of the results. We propose a novel semiparametric transformation model for such data. Compared with the existing models, our model is advantageous in the following distinctive ways. First, it allows us to provide direct estimators of the regression analysis and the association parameter. Second, the measure of surrogacy, for example, the proportion of treatment effect and relative effect, can also be directly obtained. We propose a two-stage estimation procedure for inference …

Score Test Variable Screening, Sihai Dave Zhao, Yi Li

The University of Michigan Department of Biostatistics Working Paper Series

Variable screening has emerged as a crucial first step in the analysis of high-throughput data, but existing procedures can be computationally cumbersome, difficult to justify theoretically, or inapplicable to certain types of analyses. Motivated by a high-dimensional censored quantile regression problem in multiple myeloma genomics, this paper makes three contributions. First, we establish a score test-based screening framework, which is widely applicable, extremely computationally efficient, and relatively simple to justify. Secondly, we propose a resampling-based procedure for selecting the number of variables to retain after screening according to the principle of reproducibility. Finally, we propose a new iterative score test …

A Latent Variable Transformation Model Approach For Exploring Dysphagia, Anna Snavely, David P. Harrington, Yi Li

The University of Michigan Department of Biostatistics Working Paper Series

No abstract provided.

Covariance-Enhanced Discriminant Analysis, Peirong Xu, Ji Zhu, Lixing Zhu, Yi Li

The University of Michigan Department of Biostatistics Working Paper Series

Linear discriminant analysis (LDA), a classical method in pattern recognition and machine learning, has been widely used to characterize or separate multiple classes via linear combinations of features. However, the high-dimensionality of the high-throughput features obtained from modern biological experiments, for example, microarray or proteomics, defies traditional discriminant analysis techniques. The possible interfeature correlations present additional challenges and are often under-utilized in modeling. In this paper, by incorporating the possible inter-feature correlations, we propose a Covariance-Enhanced Discriminant Analysis (CEDA) method that simultaneously and consistently selects informative features and identifies the corresponding discriminable classes. We show that, under mild regularity conditions, …