Biostatistician
Resume:
Daniel G. Polhamus
******.********@****.***
San Antonio, TX 78212
Summary
Applied statistician with an emphasis on Bayesian biostatistics, clinical trial design, and modeling of neurological counting processes. My dissertation topic is adaptive design for heterogeneous time-to-event data in a Bayesian decision theory framework.
Education
The University of Texas, San Antonio
PhD Biostatistics. In progress, dissertation topic: Frailty Models in Bayesian Clinical Design and Survival Analysis. (Advisors: Keying Ye, Nandini Kannan). Planned graduation August 2010. MS Statistics earned in May 2006.
The University of Texas, Austin
B.S. Mathematics: Probability, Statistics, and Data Analysis track. May 2003
Summary of Skills & Qualifications
Over 50 hours of graduate statistics coursework spanning Bayesian survival analysis to advanced categorical analysis
Extensive knowledge of statistical programming including Matlab, SAS (base certified 2008), R/Splus, BUGS, SPSS, JMP, MiniTab, Prism
Skilled with Mathematica, Matlab, XPPaut, Octave, WxMaxima programming
Familiar Languages/Model Tools
C, C++, Java, HTML, Pascal, SQL (and variants)
Operating Systems
Linux, Windows 95/98/XP/Vista/7/NT, MS-DOS
Experience
Research Assistant
University of Texas at San Antonio
San Antonio, TX August 2005 - Current
Assisted faculty on several projects of bio-statistical nature including:
Neuroscience (with Dr. Daijin Ko, Dr. Carlos Paladini, and Dr. Charles Wilson). Experimental design in both limited (animal response) and rich (neuronal response) data environments dealing with neuronal pharmacology and physiology. In addition, quantification of oscillatory neurological phenomena using phase resetting theory.
Step-stress (accelerated failure time) survival analysis (with Dr. Nandini Kannan); accounting for fatigue in lifetime data analysis with the specific application to pilots flying at varying altitudes.
Laser plume modeling (with Dr. Jerome Keating). With the prevalence of lasers in medicine, residue is an increasing concern. Knowledge of the statistical distribution of the residue (plume) will increase safe practice of medical and industrial lasing. This was mostly an effort in kernel density estimation.
Statistical Consulting
San Antonio, TX January 2007 - May 2008
Provided statistical advice, mentoring and analysis for many clients: sociologists (interval/Likert scale with imputation), engineers (reliability analysis), architectural (reliability analysis and 3k factorial design), and clinical scientists (data mining of the National Ambulatory Care Survey).
Mathematician I
Northrop Grumman IT
San Antonio, TX May 2003 - May 2005
Member of a 12 man modeling and simulation team to deliver verified and validated models, component libraries, applications and simulations in addition to providing statistical support to the entire division.
Developed and published a TR outlining a probabilistic model for the prediction of thermal damage risk from 1315nm laser. “A Probabilistic Model of 1315nm Laser Bio-effects”.
Provided mathematical support for the implementation of the QUEST algorithm. QUEST is an adaptive threshold estimation method using psychometric functions.
Provided modeling support to the UT Austin joint HALTing project. (Bio-effect analysis pertaining to thermal damage induced on human tissue)
Designed mixed level fractional designs and fractional factorial designs for the analysis of vision science experiments.
Provided design and analysis for the luminance bio-effect project. Also provided analysis for many vision science projects including the Four Laser experiment, Veil Glare project.
Short Courses Taught
Experimental Design and ANOVA for Biologists: Spring 2010 – An 8 hour, hands-on neurostatistics workshop presenting the classical techniques to design efficient experiments as well as the tools to analyze their results. This course covers the principles of hypothesis testing, power analysis, sample size calculations, strategies to remove undesirable sources of variability (blocks and controls) as well as commonly used experimental designs. Fixed, random, and mixed models are discussed to provide a theoretical basis to the participants.
Biostatistics and Statistical Computing for Neuroscientists: Spring 2009 – This 6 hour course served as an introduction to the R/Splus languages. Basic inference and computing based upon examples from the participating neuroscience labs was used. The modeling flexibility provided by R was emphasized in comparison to point-and-click statistical packages commonly found in the academic biology settings.
Publications
Daniel G. Polhamus, Charles J. Wilson, Carlos A. Paladini. “Truncated Estimators in Noisy PRC Estimation”, Phase Response Curves in Neuroscience. Ed. Nathan W. Schultheiss. Springer. <Under review>
Daijin Ko, Charles J. Wilson, Daniel G. Polhamus, Collin Lobb, and Carlos A. Paladini: “Identification of Bursts and Pauses in Spike Train Data”. <In preparation>
Professional Membership
The American Statistical Association
Contact this candidate