Sunday, Nov. 17 – Monday, Nov. 18; 8:30am – 5:00pm
Dr. Donald Dahlberg, Lebanon Valley College (Emeritus), Annville, PA
Dr. Neal Gallagher, Eigenvector Research, Wenatchee, WA
This is a combination of two one-day courses: Introduction to Chemometrics Without Equations (Part 1) and Multivariate Image Analysis without Equations (Part 2). A discount will be offered for the combined course over separately registering for the two one-day courses. See course descriptions for Introduction to Chemometrics Without Equations and Multivariate Image Analysis without Equations.
Part 1: Introduction to Chemometrics Without Equation is an introductory course that concentrates on two areas of chemometrics: (1) exploratory data analysis and pattern recognition, and (2) regression. Participants learn to apply techniques such as Principal Component Analysis (PCA), SIMCA, Principal Component Regression (PCR), and Partial Least Squares regression (PLS) safely. The most commonly used methods of outlier detection and data pretreatment with also be illustrated. The course emphasizes understanding the chemometric process without having to learn matrix algebra.
Part 2: Multivariate Image Analysis Without Equations (MIAwoE): Intermediate Chemometrics with a focus on hyperspectral imaging. MIAwoE is a one-day, hands-on course that builds on the popular Chemometrics Without Equations course that has been taught at EAS for over 17 years. The course emphasizes understanding the chemometric process for the expanding field of multivariate and hyperspectral image analysis without having to learn matrix algebra. Concepts that will be covered include principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA) – the most commonly used methods for image exploration, pattern recognition and outlier detection. Data pretreatment and the concept of image de-cluttering will also be discussed. Multivariate curve resolution (MCR) (a.k.a., self-modeling mixture analysis and end-member extraction) will also be introduced as a means for obtaining highly interpretable models. Finally, the concepts will be united to demonstrate the powerful concepts of target detection and targeted anomaly detection.
WHO SHOULD ATTEND
Part 1: This introductory course concentrates on two areas of chemometrics: (1) exploratory data analysis and pattern recognition, and (2) regression. Participants learn to apply techniques such as Principal Component Analysis (PCA), SIMCA, Principal Component Regression (PCR), and Partial Least Squares regression (PLS) safely. The most commonly used methods of outlier detection and data pretreatment with also be illustrated. The course emphasizes understanding the chemometric process without having to learn matrix algebra.
Part 2: This is an intermediate course for those already somewhat familiar with principal component analysis (PCA) and partial least squares (PLS) and are looking to expand their understanding into more advanced concepts with applications to multivariate / hyperspectral image analysis. Multivariate image analysis is the application of chemometric tools to hyperspectral and multispectral images used in quality assurance, astronomy, agriculture, biomedical imaging, surveillance, standoff detection, pharmaceuticals, cultural heritage, forensics and an ever growing number of fields. The course material is presented without the use of high-level mathematics with an emphasis in proper application and interpretation with a few hands on examples that allow students to explore images. (Students who wish for hands on experience will be expected to bring their own laptop and download a free 30-day demo version of Solo + MIA software available from Eigenvector Research, Inc. www.eigenvector.com)
|Day One||Day Two|
a. What is chemometrics
2. Pattern Recognition Motivation
a. What is pattern recognition
b. Relevant measurements
c. Some statistical definitions
3. Principal Component Analysis
a. What is PCA
b. Scores, loadings and eigenvalues
d. Cluster analysis
e. Mean centering and autoscaling
g. Savitzky-Golay Derivatives
a. What is regression
b. Classical least squares (CLS)
c. Inverse least squares (ILS)
d. Principal component regression (PCR)
e. Partial least squares regression (PLS)
f. Multiplicative signal correction
6. Using Regression for Pattern Recognition
a. Partial least squares – discriminant analysis
a. Review of basic chemometric techniques
include PCA, PCR, PLS
2. Multivariate Mixture Analysis
a. The probability of overlapping peaks in chromatography
b. Determining the number of components in a mixture
c. Multivariate curve resolution (MCR) and alternating least
d. Examples of time and spatial dependent systems
3. Multivariate Image Analysis (MIA)
a. Intro to 3-way arrays and simple visualizations and size/shape analyses
b. Practical multivariate image analysis (MIA)
* PCA, SIMCA, PLSDA and clustering
c. Variance Filtering for Images:
* Maximum autocorrelation factors, maximum difference factors, generalized least squares weighting (MAF, MDF, GLSW)
ABOUT THE INSTRUCTORS
Dr. Donald Dahlberg (Course Director) is Professor Emeritus of Chemistry at Lebanon Valley College. Dr. Dahlberg earned a B.S. in Chemistry from the University of Washington and a Ph.D. in Physical Chemistry from Cornell University. After decades of doing research in the area of Physical Organic Chemistry, he got involved in Chemometrics while on sabbatical in 1988 at the Center for Process Analytical Chemistry at Washington. There he learned chemometrics in the Bruce Kowalski
group (co-founder of chemometrics). Upon returning to LVC, he taught chemometrics to undergraduate students for over a decade. Although retired from the classroom, he continues to do consulting and supervises undergraduate research in industrial chemistry. Dr. Dahlberg wrote and teaches this course so that those not fluent in matrix algebra can take advantage of the powerful tool of chemometrics.
Dr. Neal B. Gallagher, PLS_Toolbox co-author and co-founder of Eigenvector Research, In., holds a doctorate in Chemical Engineering and has experience in a wide variety of applications spanning chemical process monitoring, hyperspectral image analysis, anomaly detection, quantification and classification, regression modeling and analytical instrumental development. He has extensive teaching experience including Eigenvector University and dozens of chemometric courses.