Department of Statistics Seminar
North Carolina State University

presents

Leah Martell, Marlina Dumasari & Lei Zhu

North Carolina State University

A Sampler of Graduate students

ABSTRACT

Three graduate students of the Department of Statistics at NCSU will present brief overviews of their current research.

• Leah Martell - "Wavelet Data Reduction and Multi-Resolution Decision-Making"
  Recent advances in semiconductors and sensor instrumentation provide opportunities for practitioners to collect data, useful in detecting problems in operations, such as manufacturing productions, image monitoring for security checking, unusual behavior in telephone usages and so on. With automatic data collection tools, the data size becomes too large for the standard statistical procedures to handle at the regular data scale, especially on real-time operations. This presentation proposes the idea of analyzing the "compressed data" to detect process problems. We focus on the study of model selections om wavelet data compression. This talk presents and compares several criteria of selecting wavelet coefficients (used as the "compressed data" in our study) to balance the goodness-of-fit and data compression requirements. We illustrate the proposed ideas with several real-life examples of detecting faults in manufacturing processes. Our presentation will be conducted by pointing out a few directions for future research.

• Marlina D. Nasution - "Joint Analysis of Resighting and Radio-Telemetry Data"
  Biologists often use more than one marking technique in wildlife studies. From each of the mark types, it is common to conduct a separate analysis of the recapture data to estimate parameters of interest, such as survival rates. Two data types that can be used in estimating survival rates are resighting and radio--telemetry data. The Jolly--Seber model is commonly used to analyze the resighting data, while the Kaplan--Meier product limit estimators, modified for staggered entry of animals, are used to analyze the radio--telemetry data. In a study where some birds receive two types of tags (i.e., leg bands and radios) and other birds receive just one tag type (leg bands), the separate Jolly--Seber and Kaplan--Meier analyses do not exploit all of the information in the combined data sets.
  In this paper, we propose a model and likelihood for the combined analysis of resighting and radio--telemetry data. In comparison to the separate analyses, this richer model provides more information about the biology and sampling processes. For example, the richer model permits assesment of assumptions required by the separate analyses and allows estimation of additional parameters. As an example, we apply the model to annual resighting and monthly telemetry data from a population of Snail Kites.

• Lei Zhu - "Statistical Decoding of Potent Pools Based on Chemical Structure"
  Pooling experiments are widely used as a cost-effective approach for screening chemical compounds as part of the drug discovery process in pharmaceutical companies. When a biologically potent pool is found, the goal is to decode the pool, that is, to determine which of the individual compounds are potent. Usually, only the responses are used in making the determination. But it is known that biological potency of a compound is highly dependent on its chemical structure. Thus we use compound structure information, in terms of atom pairs, as covariates for predicting the potency of individual compounds. Covariate effects are estimated in models with and without retesting. The procedure is demonstrated using real data from a drug discovery process at GlaxoWellcome.

Friday, October 29, 1999

3:35 - 4:35 pm

206 Cox Hall

Refreshments will be served on the second floor of Dabney Hall (left of Room 222) at 3:00 pm.