作者: Wilm Schumacher , Stephan Stöckel , Petra Rösch , Jürgen Popp
DOI: 10.1002/JRS.4568
关键词: Independent component analysis 、 Bacillus weihenstephanensis 、 Linear discriminant analysis 、 Bacillus mycoides 、 Analytical chemistry 、 Dimensionality reduction 、 Bacillus licheniformis 、 Pattern recognition 、 Mathematics 、 Artificial intelligence 、 Chemometrics 、 Principal component analysis
摘要: In this contribution a new method for improving the accuracy of classification and identification experiments is presented. For purpose four most applied dimension reduction methods (principal component analysis, independent partial least square linear discriminant analysis) are used as starting point optimization method. The done by specially designed genetic algorithm, which best suited kind experiments. presented multi-level chemometric approach has been tested Raman dataset containing over 2200 spectra eight classes bacteria species (Bacillus anthracis, Bacillus cereus, licheniformis, mycoides, subtilis, thuringiensis, weihenstephanensis Paenibacillus polymyxa). improved 6% compared with accuracy, if standard applied. rate 14% reduction. testing in experiment showed robustness algorithm. Copyright © 2014 John Wiley & Sons, Ltd.