Biography
I am primarilly interested in Bayesian Statistics. Within this rubrick I am working on:
a) Biostatistics: 1) Sample Size Determination (in a clinical trial) 2) Predicting the Final Event (in a clinical Trial) and 3) Disease Modification
b) Recurrent Neural Networks and Bayesian Word2Vec Models
c) Hidden Markov and Particle Filter Model
d) Dependent Binary Random fields (with Garry witt)
My student Sam Ackerman got his PhD in April using Hidden Markov and Particle Filter Models
My student Lucas Glass is currently writing a thesis on Recurrent Neural Networks and Bayesian Word2Vec Models
Research Interests
- Clinical Trials: (i) Bayesian sample size re-estimation and (ii) design Word2Vec: A Bayesian Probabilistic Disease model for time-stamped Medical Data. (Work joint with Lucas Glass) Sequential Markov Chain Monte Carlo Techniques: Particle Filters and GPS analysis (Work joint with Samuel Ackerman) Multiresolution Markov Random Fields. Correlated Binary Random Fields.
Courses Taught
Number | Name | Level |
|---|---|---|
STAT 1001 | Quantitative Methods for Business I | Undergraduate |
Selected Publications
Recent
Turkoz, I., Sobel, M., & Alphs, L. (2019). Application of Bayesian analyses to doubly randomized delayed start, matched control designs to demonstrate disease modification. Pharmaceutical Statistics, 18(1), 22-38. doi: 10.1002/pst.1905.
Sobel, M. & Turkoz, I. (2018). Bayesian blinded sample size re-estimation. Communications in Statistics - Theory and Methods, 47(24), 5916-5933. doi: 10.1080/03610926.2017.1404097.
Sobel, M. & Turkoz, I. (2021). Predicting the timing of the final event in a clinical trial using spline process simulations. jsm, Chicago, IL.