Physical Sciences and Mathematics Commons^™

Quantile Regression For Survival Data With Delayed Entry, Boqin Sun

Doctoral Dissertations

Delayed entry arises frequently in follow-up studies for survival outcomes, where additional study subjects enter during the study period. We propose a quantile regression model to analyze survival data subject to delayed entry and right-censoring. Such a model offers flexibility in assessing covariate effects on survival outcome and the regression coefficients are interpretable as direct effects on the event time. Under the conditional independent censoring assumption, we proposed a weighted martingale-based estimating equation, and formulated the solution finding as a $\ell_1$-type convex optimization problem, which was solved through a linear programming algorithm. We established uniform consistency and weak convergence of …

Variational Approximations For Density Deconvolution, Yue Chang

Doctoral Dissertations

This thesis considers the problem of density estimation when the variables of interest are subject to measurement error. The measurement error is assumed to be additive and homoscedastic. We specify the density of interest by a Dirichlet Process Mixture Model and establish variational approximation approaches to the density deconvolution problem. Gaussian and Laplacian error distributions are considered, which are representatives of supersmooth and ordinary smooth distributions, respectively. We develop two variational approximation algorithms for Gaussian error deconvolution and one variational approximation algorithm for Laplacian error deconvolution. Their performances are compared to deconvoluting kernels and Monte Carlo Markov Chain method by …

Model-Based Predictive Analytics For Additive And Smart Manufacturing, Zhuo Yang

Doctoral Dissertations

Qualification and certification for additive and smart manufacturing systems can be uncertain and very costly. Using available historical data can mitigate some costs of producing and testing sample parts. However, use of such data lacks the flexibility to represent specific new problems which decreases predictive accuracy and efficiency. To address these compelling needs, in this dissertation modeling techniques are introduced that can proactively estimate results expected from additive and smart manufacturing processes swiftly and with practical levels of accuracy and reliability. More specifically, this research addresses the current challenges and limitations posed by use of available data and the high …

Essays In Financial Economics: Announcement Effects In Fixed Income Markets, James J. Forest

Doctoral Dissertations

ABSTRACT ESSAYS IN FINANCIAL ECONOMICS: ANNOUNCEMENT EFFECTS IN FIXED INCOME MARKETS PHD IN FINANCE MAY 2018 JAMES J FOREST B.A., FRAMINGHAM STATE UNIVERSITY M.S., NORTHEASTERN UNIVERSITY Ph.D., UNIVERSITY OF MASSACHUSETTS – AMHERST Directed by: Professor Hossein B. Kazemi This dissertation demonstrates the use of empirical techniques for dealing with modeling issues that arise when analyzing announcement effects in fixed income markets. It describes empirical challenges in achieving unbiased and efficient parameter estimates and shows the importance of modelling a wide range of macroeconomic announcement effects to avoid omitted variable bias. Employing techniques common in Macroeconomics, financial market researchers are better …

Asymptotic Behavior Of The Random Logistic Model And Of Parallel Bayesian Logspline Density Estimators, Konstandinos Kotsiopoulos

Doctoral Dissertations

This dissertation is comprised of two separate projects. The first concerns a Markov chain called the Random Logistic Model. For r in (0,4] and x in [0,1] the logistic map f_r(x) = rx(1 - x) defines, for positive integer t, the dynamical system x_r(t + 1) = f(x_r(t)) on [0,1], where x_r(1) = x. The interplay between this dynamical system and the Markov chain x_r,N(t) defined by perturbing the logistic map by truncated Gaussian noise scaled by N^-1/2, where N -> infinity, is studied. A natural question is …

On Modeling Quantities For Insurer Solvency Against Catastrophe Under Some Markovian Assumptions, Daniel Jefferson Geiger

Doctoral Dissertations

"Insurance companies sometimes face catastrophic losses, yet they must remain solvent enough to meet the legal obligation of covering all claims. Catastrophes can result in large damages to the policyholders, causing the arrival of numerous claims to insurance companies at once. Furthermore, the severity of an event could impact the time until the next occurrence. An insurer needs certain levels of startup capital to meet all claims, and then must have adequate reserves on a continual basis, even more so when catastrophes occur. This work examines two facets of these matters: for an infinite time horizon, we extend and develop …

New Developments Of Dimension Reduction, Lei Huo

Doctoral Dissertations

"Variable selection becomes more crucial than before, since high dimensional data are frequently seen in many research areas. Many model-based variable selection methods have been developed. However, the performance might be poor when the model is mis-specified. Sufficient dimension reduction (SDR, Li 1991; Cook 1998) provides a general framework for model-free variable selection methods.

In this thesis, we first propose a novel model-free variable selection method to deal with multi-population data by incorporating the grouping information. Theoretical properties of our proposed method are also presented. Simulation studies show that our new method significantly improves the selection performance compared with those …