Thesis (Ph. D.)--University of Chicago, Dept. of Statistics, August 2001. / Includes bibliographical references. Also available on the Internet.

Thesis (Ph. D.)--University of Texas at Austin, 2002. / Vita. Includes bibliographical references (leaves 191-214). Available also in a digital version from Dissertation Abstracts.

Thesis (Ph. D.)--University of Washington, 2003. / Vita. Includes bibliographical references (leaves 61-65).

Value of information (VOI) is the amount a decision maker is willing to pay for information to better understand the uncertainty surrounding a decision, prior to making the decision. VOI is a key part of decision analysis (DA). Especially in this age of information explosion, evaluating information value is critical. VOI research tries to derive generic conclusions regarding VOI properties. However, in most cases, VOI properties rely on the specific decision context, which means that VOI properties may not be generalizable. Thus, instead, VOI properties have been derived for typical or representative decisions. In addition, VOI analysis as a method of DA has been successfully applied to practical decision problems in a variety of industries. This approach has also been adopted as the basis of a heuristic algorithm in the latest research in simulation and optimization. Portfolio Decision Analysis (PDA), rooted in DA, is a body of theories, methods, and practices that seek to help decision makers with limited budget select a subset of candidate items through mathematical modeling that accounts for relevant constraints, preferences, and uncertainties. As one of the main tools for resource allocation problems, its successful implementation, especially in capital-intensive industries such as pharmaceuticals and oil & gas, has been documented (Salo, Keisler and Morton 2011). Although VOI and PDA have been extensively researched separately, their combination has received attention only recently. Resource allocation problems are ubiquitous. Although significant attention has been directed at it, less energy has been focused on understanding the VOI within this setting, and the role of VOI analysis to solve resource allocation problems. This belief motivates the present work. We investigate VOI properties in portfolio contexts that can be modeled as a knapsack problem. By further looking at the properties, we illustrate how VOI analysis can derive portfolio management insights to facilitate PDA process. We also develop a method to evaluate the VOI of information portfolios and how the VOI will be affected by the correlations between information sources. Last, we investigate the performance of a widely implemented portfolio selection approach, the benefit-cost ratio (BCR) approach, in PDA practice. / text

Failure time data analysis, or survival analysis, is involved in various research fields, such as medicine and public health. One basic assumption in standard survival analysis is that every individual in the study population will eventually experience the event of interest. However, this assumption is usually violated in practice, for example when the variable of interest is the time to relapse of a curable disease resulting in the existence of long-term survivors. Also, presence of unobservable risk factors in the group of susceptible individuals may introduce heterogeneity to the population, which is not properly addressed in standard survival models. Moreover, the individuals in the population may be grouped in clusters, where there are associations among observations from a cluster. There are methodologies in the literature to address each of these problems, but there is yet no natural and satisfactory way to accommodate the coexistence of a non-susceptible group and the heterogeneity in the susceptible group under a univariate setting. Also, various kinds of associations among survival data with a cure are not properly accommodated. To address the above-mentioned problems, a class of models is introduced to model univariate and multivariate data with long-term survivors. A semiparametric cure model for univariate failure time data with long-term survivors is introduced. It accommodates a proportion of non-susceptible individuals and the heterogeneity in the susceptible group using a compound- Poisson distributed random effect term, which is commonly called a frailty. It is a frailty-Cox model which does not place any parametric assumption on the baseline hazard function. An estimation method using multiple imputation is proposed for right-censored data, and the method is naturally extended to accommodate interval-censored data. The univariate cure model is extended to a multivariate setting by introducing correlations among the compound- Poisson frailties for individuals from the same cluster. This multivariate cure model is similar to a shared frailty model where the degree of association among each pair of observations in a cluster is the same. The model is further extended to accommodate repeated measurements from a single individual leading to serially correlated observations. Similar estimation methods using multiple imputation are developed for the multivariate models. The univariate model is applied to a breast cancer data and the multivariate models are applied to the hypobaric decompression sickness data from National Aeronautics and Space Administration, although the methodologies are applicable to a wide range of data sets. / published_or_final_version / Statistics and Actuarial Science / Master / Master of Philosophy

Capillary-based analytical methodologies were developed to meet the need for metabolite determination in two major areas. The first area is the determination of free bilirubin in sera for the management of jaundiced neonates under critical conditions. Three sensitive detection techniques were investigated, Quantum dots (QD) mediated fluorescence, Chemiluminescence (CL) and Microelectrode detection. Four different types of QDs were synthesized for the direct bilirubin determination. The CAH-capped CdTe QDs were selected as it shows the best performance compared to organic dyes and other QDs. Its optimized preparation conditions are: refluxing solution containing Cd/Te/CAH (1:0.5:2.4 w/w) for 4 hours at 100 °C. From Transmission Electron Microscope characterization, nano-size QDs with an uniform size distribution, high luminescence and good stability were obtained. The optimized detection conditions were: incubation of bilirubin with CAH-capped CdTe QDs (5 10-6 mol/L) in water at pH=5.6 and 20 oC for 8 min prior to spectrofluorometric determination (λex=473 nm and λem=580 nm). A linear working range from 0.043-0.86 μg/mL with 0.9943 correlation coefficient and 2 ng/mL detection limit (LOD, S/N=3) were achieved. Results from nFIA-CL indicate a quick response within seconds though a poorer LOD (S/N=3) of 15 μg/mL for the direct bilirubin determination. The third technique investigated used an enzyme microelectrode and it was found to be able to couple with capillary electrophoresis (CE) in frontal analysis (FA) for the determination of free bilirubin in serum samples. Making use of the micron size of the carbon-fiber electrode, a new MCNTs (Multi-wall Carbon Nanotubes) modified CFMEs (Carbon fiber microelectrodes) was fabricated within a microchip-CE device with three guided channels to enable electrodes alignment. Method to immobilize bilirubin oxidase (BOD) onto the CFMEs surface by the carbodiimide chemistry achieved the highest detection sensitivity. Under optimized conditions (sample introduced by hydrodynamic injection at △H (20 cm), and a running/detection buffer (10 mM phosphate) at pH 7.4, working potential for amperometric detection at +0.8 V), a linear working range between 1-40 μg/mL and a detection limit (S/N=3) at 0.15 μg/mL for free bilirubin was achieved. The second area for metabolite determination was developing a new analytical method for the management of APL (acute promyelocytic leukemia) patients under arsenic treatment, a drug required continued monitoring. The analytical requirements include a high detection sensitivity and the capability to provide timely results for multiple drug residues. Using a 20 mM phosphate as the running buffer and 0.05mM CTAH (Cetyl-trimethyl-ammonium hydroxide) as an additive for EOF reversal, co-EOF (co-electroosmotic flow) stacking was established to enhance up to 200 times of the detection limit for arsenite. Satisfactory baseline separation for arsenite, arsenate, MMA (Methylarsonic acid) and DMA (Dimethylarsinic acid) was achieved with linear working ranges (correlation coefficients > 0.999) from 1-50 μg/mL for arsenate and DMA, 0.5-50 μg/mL for MMA as well as 0.1-50 μg/mL for arsenite. Detection limits (S/N=3, n=3) achievable for arsenate, arsenite, MMA and DMA were found to be 0.41 μg/mL, 0.01 μg/mL, 0.04 μg/mL and 0.32 μg/mL respectively at levels meeting the requirement for APL patient urine monitoring. / published_or_final_version / Chemistry / Doctoral / Doctor of Philosophy

Not available / text

Not available / text

Not available / text

Not available / text