The seminar will take place on Fridays at 2:00pm in the Business Building (BB) room 3.02.30 (unless stated otherwise). See campus map. For additional information contact Victor De Oliveira, (210) 458-6592.

February 15, 2008 (in BB 4.02.10, Executive Room)

• David Han, Department of Mathematics and Statistics, McMaster University
• Title: Step-stress Tests and Optimal Progressive Type-I Censoring
  • Abstract: In reliability studies, accelerated life-testing allows for gradual increment of stress levels on test units during an experiment. In a special class of accelerated life tests known as step-stress tests, the stress levels increase discretely at pre-fixed time points, and this allows the experimenter to obtain information on the parameters of the lifetime distributions more quickly than under normal operating conditions. In this talk, a k-level step-stress accelerated life-testing is considered with an equal step duration tau. In particular, the case of progressively Type-I right censored data from an exponential distribution is investigated with a single stress variable. For small to moderate sample sizes, a practical modification is suggested to guarantee a feasible k-level step-stress test under progressive Type-I censoring, and the optimal tau is determined under C-, D-, and A-optimality criteria. Next, we discuss the determination of optimal tau Ï„ under the condition that the step-stress test proceeds to the k-th stress level, and the efficiency of this conditional inference is compared to that of the unconditional case.

February 22, 2008 (in BB 4.02.10, Executive Room)

• Xuefeng Liu, Department of Internal Medicine, Wayne State University
• Title: Joint Models for the Association of Longitudinal Binary and Continuous Processes with Application to a Smoking Cessation Trial
  • Abstract: Joint models for the association of a longitudinal binary and a longitudinal continuous process are proposed for situations where their association is of direct interest. The models are parameterized such that the dependence between the two processes is characterized by unconstrained regression coefficients. Bayesian variable selection techniques are used to parsimoniously model these coefficients. An MCMC sampling algorithm is developed for sampling from the posterior distribution, using data augmentation steps to handle missing data. Several technical issues are addressed to implement the MCMC algorithm efficiently. The models are motivated by, and are used for, the analysis of a smoking cessation clinical trial in which an important question of interest was the effect of the (exercise) treatment on the relationship between smoking cessation and weight gain.

February 29, 2008

• Changxiang Rui, Department of Mathematical Sciences, University of Arkansas
• Title: On Optimal Prediction in Log-Gaussian Random Fields
  • Abstract: In this talk I consider the problems of point and block (integral)prediction in log-Gaussian random fields for the case when the mean of the log-process is not constant and depends linearly on unknown parameters. First, I propose a new point predictor that is optimal within a certain family of predictors, which extend a result in De Oliveira (2006) that holds in the case when the mean of the log-process is constant; furthermore, I derive a point prediction interval which is the shortest (optimal) within a certain class of prediction intervals. Second, I show that the results in De Oliveira (2006) regarding optimal block prediction cannot be extended to the case when the mean of the log-process is not constant. Specifically, it is shown that the two families of block predictors considered by De Oliveira lack an optimal predictor. Finally, I numerically compare the predictive efficiency of the proposed point and block predictors, and a numerical comparison between the derived shortest prediction interval and a naive prediction interval is provided.

March 7, 2008

• Kihoon Yoon, Department of Epidemiology and Biostatistics, University of Texas Health Science Center at San Antonio
• Title: Over-represented sequences located on UTRs are potentially involved in regulatory functions
  • Abstract: Eukaryotic gene expression must be coordinated for the proper functioning of biological processes. This coordination can be achieved both at the transcriptional and post-transcriptional levels. In both cases, regulatory sequences placed at either promoter regions or on UTRs function as markers recognized by regulators that can then activate or repress different groups of genes according to necessity. While regulatory sequences involved in transcription are quite well documented, there is a lack of information on sequence elements involved in post-transcriptional regulation. We used a statistical over-representation method to identify putative regulatory elements located on UTRs. An exhaustive search approach was used to calculate the frequency of all possible n-mers (short nucleotide sequences) in 16,160 human genes of NCBI RefSeq sequences and to identify any peculiar usage of n-mers on UTRs. After a stringent filtering process, we identified circa 4,000 highly over-represented n-mers on UTRs. We provide evidence that these n-mers are potentially involved in regulatory functions. Identified n-mers overlap with previously identified binding sites for HuR and Tia1 and, AU-rich sequences. We determined also that over-represented n-mers are particularly enriched in a group of 159 oncogenes. Finally, a method to cluster n-mer groups allowed the identification of putative gene networks.

April 11, 2008

• Zhiming Yang, Information and Assessment, Pearson
• Title: IRT and Its Application in Clinical Assessment
  • Abstract: Item Response Theory (IRT) is a set of probabilistic models that tries to describe the relationship between an examinees test performance and the latent trait underlying performance. Latent traits may be cognitive ability, achievement, personality, affective commitment, etc. There are many IRT models. The commonly used models are the one-parameter logistic model (Rasch model), two-parameter logistic model, three-parameter model, and partial credit or graded response model. IRT is the theoretical foundation of modern Computerized Adaptive Testing (CAT). It was widely used in the testing business with tests such as the GRE, GMAT, PCAT, Standford Achievement Test, Differential Ability Scales (DAS), Wechsler Intelligence Scales (WISC-IV, WAIS-III), and Minnesota Multiphasic Personality Inventory (MMPI). The advantages of IRT include getting "sample-free" item calibrations and "item-free" person measurement, providing a better framework for equating and vertical scaling, and controlling the standard error of a selected test at any desired set of score levels. Additional advantages include providing new methods for standard setting, saving testing time while keeping or improving the quality of a test, and providing more information than Classical Testing Theory (CTT). This seminar will introduce some basic concepts of IRT and its application in clinical assessment. Some challenges for statisticians will be also discussed.

April 18, 2008

• Robert Lieli, Department of Economics, University of Texas at Austin
• Title: Prediction of Binary Outcomes
  • Abstract: We address the issue of using a set of covariates to categorize or predict a binary outcome. This is a common problem in many disciplines including economics. In the context of a prespecified utility (or cost) function we examine the construction of forecasts suggesting an extension of the Manski (1975, 1985) maximum score approach. We provide analytical properties of the method and compare it to more common approaches such as forecasts or classifications based on conditional probability models. The results are informative for both forecasting environments as well as program allocation where the value of including the participant in the program depends on how useful the program turns out to be for that participant.

Past Seminars: Fall 2007  Spring 2007  Fall 2006