Previous Topic Back Forward Next Topic
Print Page Dr. Frank Dieterle
 
Ph. D. ThesisPh. D. Thesis 6. Results  Multivariate Calibrations6. Results Multivariate Calibrations
Home
News
About Me
Ph. D. Thesis
  Abstract
  Table of Contents
  1. Introduction
  2. Theory Fundamentals of the Multivariate Data Analysis
  3. Theory Quantification of the Refrigerants R22 and R134a: Part I
  4. Experiments, Setups and Data Sets
  5. Results Kinetic Measurements
  6. Results Multivariate Calibrations
    6.1. PLS Calibration
    6.2. Box-Cox Transformation + PLS
    6.3. INLR
    6.4. QPLS
    6.5. CART
    6.6. Model Trees
    6.7. MARS
    6.8. Neural Networks
    6.9. PCA-NN
    6.10. Neural Networks and Pruning
    6.11. Conclusions
  7. Results Genetic Algorithm Framework
  8. Results Growing Neural Network Framework
  9. Results All Data Sets
  10. Results Various Aspects of the Frameworks and Measurements
  11. Summary and Outlook
  12. References
  13. Acknowledgements
Publications
Research Tutorials
Links
Contact
Search
Site Map
Guestbook
Print this Page Print this Page

6.   Results Multivariate Calibrations

In this chapter, the data sets of the refrigerants R22 and R134a, which were introduced in section 4.5.1.1, are investigated by the use of the most common methods of multivariate calibration starting with the PLS. Thereby models for the relationship between the 40 time-resolved sensor responses and the concentrations of both analytes are established. As the linear PLS calibration cannot deal with the nonlinearities present in the data sets, several methods, which are known to be capable of dealing with nonlinearities, are applied to this data set afterwards. These methods originate from different fields of scientific research such as multivariate spectroscopic calibration, quantitative structure activity relationship, machine learning, medical decision support systems, psychometrics, economic research and artificial intelligence whereby all these methods are able to calibrate multivariate relationships. An overview of the prediction errors for the calibration data and the validation data is shown in table 2 in section 6.11 for all methods used in this chapter. It is obvious that the calibration quality of the different methods shows a very broad variety ranging from unacceptable results for the widely used PLS calibration to promising results for neural network based calibrations.

Page 68 © Dr. Frank Dieterle, 14.08.2006 Navigation