Data Analysis Made Easy
These pages are provided to give you a head start into various topics covered by our courses.
Do More with Less: Create Smart Data with Design of Experiments!
Research and development means:
- Data often still has to be created
- Creating data is costly
Why do many industries not optimize the ROI on this data investment?
It should be clear that sub-optimal data in R&D with a high probability leads to sub-optimal processes or products both in terms of quality and performance.
The science of systematically creating "optimal" data in the most efficient manner is design of experiments.
Omics and Biomarker Study Design
Omics and biomarker studies are inherently comparative. - We compare healthy to diseased, responders to nonresponders, good versus poor prognosis. The focus is usually on the how of this comparison (technology), but it matters just as much what we compare (samples, conditions). Whereas it is generally very difficult and costly to improve the measurement performance, it is often rather simple to achieve a quantum leap in result quality and validity by an improved study design.
SVM Starter
Support Vector Machines and related Kernel methods are a very versatile and flexible approach rooted in statistical learning theory. Thus, they address one of the most critical issues in applied data analysis: Deriving a generalizable, valid model for future data from a limited amount of valuable training examples.
Data Fusion
Data fusion is about a systematic, joint analysis of data from different sources (measurement approaches) aiming to harvest their synergy, trying to unravel the big picture. As rarely any single data source is able to capture the full information, there is a great current interest in this subject from various fields ranging from bioinformatics, chemometrics, pattern recognition to engineering.