Our group uses a combination of mathematical and statistical approaches to develop novel algorithms and methods for analyzing chemical data, a field of study known as chemometrics. In other words, it combines chemistry and statistics to extract as much relevant information as possible from chemical data. This approach can be used to build calibration models, predicting a property of interest, or classification models, predicting the identity of a species. Some of these principles include signal processing, multivariate statistics, partial least squares regression, multivariate curve resolution and more. We apply these techniques to a number of or projects, including the analysis of edible oils, planetary/martian geological samples, and provenance of rosewood.
Publications
- Ottaway, J.; Smith, J.P.; Booksh, K.S. “Adaptive Regression via Subspace Elimination” ACS Symposium Series 1199: 40 Years of Chemometrics – From Bruce Kowalski to the Future 2015 241-256
