Statistical methods to estimate physical activity. Statistical methodology development: measurement error correction methods, non-parametric / statistical learning methods, Bayesian data analysis, smoothing
I am statistician, and my students and I have collaborated extensively with researchers from kinesiology and public health to develop statistical methods to estimate various aspects of physical activity from accelerometers and other devices. We have also developed novel statistical methodology for measurement error correction, non-parametric / statistical learning, and Bayesian data analysis.
Recent and current projects include:
- Free-living physical activity and sedentary behavior: novel accelerometer data processing methods
- Physical activity calibration in individuals with movement limitations
- Dimension reduction methods for hierarchical functional data
- Calibration and measurement error in physical activity recall surveys and accelerometers
- Behavioral database architecture for the storage, analysis, and reporting of biosignal datasets.
- Use of objective physical activity monitors and accelerometers in clinical trials
- Transdisciplinary research on energetics and cancer
- BA Williams College, 1992
- PhD Cornell University, 2000
- Postdoctoral Training: Harvard School of Public Health, 2000-2001