Support Vector Machines in Pharmaceutical Analysis (FWOSL31)

Identification of pharmaceuticals and determination of the content of pharmaceutical components are the fundamentals in pharmaceutical analytical technology. With the technological development in analytical instruments, it has become important to obtain the pharmaceutical composition and content by analyzing the signals acquired from analytical instruments. Automatic signal recognition and multi-component analysis are therefore crucial to pharmaceutical analysis. The proposed research plan will evaluate the applicability of the chemometric technique Support Vector Machines in those contexts. In the first context the classification of plant extracts based on chromatographic fingerprints will be evaluated. In the second context, the data analysis of chromatographic fingerprints of plant extracts will be studied.
Support vector machine (SVM) was introduced by Vapnik according to the statistical learning theory. It is a novel powerful learning method based on the Structural Risk Minimization (SRM) principle. This learning method is gaining popularity due to many attractive features. It can solve both linear and non-linear problems in regression, pattern recognition and classifications effectively. At the moment, the currently applied chemometric techniques in the above contexts are often not capable of obtaining good data analysis results. SVM has both theoretically as practically a great potential to obtain better analysis results.

Keywords:Drug Analysis, Chemometrics, Analytical Chemistry, Chromatography, Artificial Intelligence, Food Analysis

Disciplines:Pharmaceutical sciences, Chemical sciences