|
|
|
|
|
|
Professor Lennart Eriksson, born 1963, has a Ph.D. in Chemometrics and Organic Chemistry from Umeå University, Sweden, where he has been Associate Professor in Chemometrics since 1995. His research has concerned multivariate statistical techniques (PCA & PLS) as applied to QSAR in drug design and environmental sciences. More recently, his research interests have included extensions of the basic methods (OPLS, non-linear PLS, hierarchical PLS, time-resolved QSAR,…) and their applicability to the complex data mining and data integration problems encountered within the life science area (e.g., the ‘omics’, ‘informatics’ and drug design fields). Lennart Eriksson has authored and co-authored more than 80 publications. He is since 1994 employed by Umetrics AB as senior lecturer and application specialist. Umetrics AB, a member of the MKS Instrument Group, develops software for multivariate data analysis and design of experiments.
|
| |
|
|
| Workshop, 14.30-17.30, 5 July & 09.00-10.30 6 July 2006, Chemistry Research Laboratory, Oxford University |
|
Data Analysis & Visualisation in Discovery Chemistry & Biology
Workshop Instructor: Lennart Eriksson, Umetrics
The workshop will focus on projection methods such as Principal Components Analysis (PCA), Partial Least Square (PLS) and the newly developed Orthogonal-PLS (OPLS).
Examples from discovery chemistry and discovery biology will be used to exemplify how the following problem areas can be addressed:
*Large data sets (data bases) many records, cases, observations, and possibly also many variables, indicators, features, etc;
*Secondary analysis the data were collected for other objectives than the one underlying the present analysis;
*Heterogeneous data the data are clustered in unknown ways, and relationships between variables may change between clusters;
*Non-independence both observations and variables are often dependent and/or correlated, but in different ways and degrees in different parts of the data;
*Selection bias different categories of cases have different amounts of data recorded;
*Drift in the data data measured at different times have different means and variances and relationships;
*Non-numeric data qualitative variables, or just pieces of unorganized text are often mixed with quantitative variables.
Participants will also take home:
1) A personal copy of two advanced textbooks optimal for continued learning and application:
a) Multi- and Megavariate Data Analysis Part I: Basic Principles and Applications, L. Eriksson, E. Johansson, N. Kettaneh-Wold, J.Trygg, C. Wikström, and S. Wold, and b) Multi- and Megavariate Data Analysis Part II: Advanced Applications and Method Extensions, L. Eriksson, E. Johansson, N. Kettaneh-Wold, J.Trygg, C. Wikström, and S. Wold.
2) a CD with demonstration versions of the Umetrics software.
|
|
|
|
|
|
|
