MULTIVAR, PLANEX, INTERLAB — From a Collection of Algorithms to an Expert System: Statistics Software Written from Chemists for Chemists

  • D. Wienke
  • U. Wank
  • M. Wagner
  • K. Danzer
Conference paper


Three software packages for specialized statistical analysis in the chemical laboratory are presented.

The first system MULTIVAR was developed between 1985 and 1988 as menu driven software package for multivariate statistical data analysis containing a collection of the most important algorithms for data scaling, factor analysis, multivariate regression and pattern recognition applications. An on-line help about every algorithm simplifies the use of the program for the chemist.

The second Program PLANEX was created between 1988 and 1990 as a research study about expert systems for method choice in statistics.

PLANEX realizes a dialog with the chemist about the mathematical speciality of optimal experimental designs. After that dialog generates PLANEX an optimal experimental design as a kind of laboratory worksheet on printer. The most optimal experimental plan is selected using a so called decision tree.

This support by a decision tree represented a higher level of software intelligence in comparison to MULTIVAR as a Poor collection of statistical algorithms.

A further level of software intelligence was implemented in the third program INTERLAB — an expert system for the evaluation of interlaboratory comparisons.

INTERLAB analyses stepwise and automatically given experimental data and offers at every stage of the analyses the optimal method from a pool of possible parametric and robust statistical tests, As result of the combination of this automatic data evaluation with a user — program — dialog as an overlay Process an exhaustive statistical analysis of interlaboratory test data is available.


Partial Little Square Expert System Anal Chim Optimal Experimental Design Software Intelligence 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Goldstein M, Dillon WR (1984) Multivariate Analysis, Methods and Applications, John Wiley and Sons, New York- Chichester-Brisbane-Toronto-SingaporeGoogle Scholar
  2. 2.
    Fahrmeir L, Hamerle A (1984) Multivariate Statistische Verfahren, Walter de Gruyter, Berlin - New YorkGoogle Scholar
  3. 3.
    Varmuza, K (1978) Lecture Notes in Chemistry - Pattern Recognition in Chemistry, Springer-Verlag, Berlin-Heidelberg-New YorkGoogle Scholar
  4. 4.
    Sharaf MA, Illman DL, Kowalski BR (1986) Chemometrics, Wiley, New York-Chichester-Brisbane-Toronto-SingaporeGoogle Scholar
  5. 5.
    Massart DL, Vandeginste BGM, Deming SN, Michotte Y, Kaufman L (1987), Chemometrics - a textbook, Elsevier Science Publishers, Amsterdam - New YorkGoogle Scholar
  6. 6.
    SPSSPC - Statistical Software Package for Social Sciences, (1985), SPSS Inc, User ManualGoogle Scholar
  7. 7.
    SAS Institute Inc (1982), SAS User’s Guide, Statistics, NC, SAS Institute, IncGoogle Scholar
  8. 8.
    STATGRAPHICS - software package (1985) Statistical Graphics Corporation, User’s Manual, Graphic Software Systems Inc, Trademark of Statistical Graphics CorpGoogle Scholar
  9. 9.
    Wold S, Sjöström M, Carlson R, Lundstedt T, Hellberg S, Skagerberg B, Wikström C, Öhman J (1986) Multivariate Design, Anal Chim Acta 191:17CrossRefGoogle Scholar
  10. 10.
    Wienke D, Jana A, Mikkeleit W, Danzer K (1988) Joint Calibration of Analyte and Disturbing Component instead of Their Separation, Vortrag auf der 4 Tagung Computereinsatz in der Analytischen Chemie (COMPANA”88), Jena, TagungsbandGoogle Scholar
  11. 11.
    Wienke D, Gitter M, Aures J, Byhan HG, Münch U, Pohl HJ, Danzer K (1989) Anal Chim Acta 223:247CrossRefGoogle Scholar
  12. 12.
    Wienke D, Einax J, Baltes U, Danzer K (1991) Treatment of a class-in-class modelling problem in chemical pattern recognition, Anal Chim Acta, in printGoogle Scholar
  13. 13.
    Wienke D, Danzer K (1990) Application of the k-th Nearest Neighbour Pattern Recognition Method in Chemical Polyoptimization (Optative Vector Method), EUROANALYSIS CONFERENCE 1990 / COBAC 5, ViennaGoogle Scholar
  14. 14.
    Wienke D, Gitter M, Danzer K (1990) Glass Design by Multivariate Methods - the Software Package MULTIVAR for the Development of Heliovar SB, Wiss Zeitschrift der FSU Jena (Proceedings 4th International Otto-Schott-Coliogium of Glas Chemistry)Google Scholar
  15. 15.
    Wienke D, Pflug G, Danzer K, Dunken H (1987) Multivariate Techniques for Glas Structure Analysis, 5th Conference Solid State Analysis, TU Karl-Marx-Stadt, proceedingsGoogle Scholar
  16. 16.
    Bandemer H, Näther W (1980) Theorie und Anwendung der optimalen Versuchsplanung, Band I-II, Akademie-Verlag, BerlinGoogle Scholar
  17. 17.
    Scheffler E (1984) Einführung in die Praxis der statistischen Versuchsplanung, VEB Deutscher Verlag für Grundstoffindustrie, LeipzigGoogle Scholar
  18. 18.
    Hartmann K, Lezki E, Schäfer W (1988) Statistische Versuchsplanung und Auswertung in der Stoffwirtschaft, VEB Deutscher Verlas für Grundstoffindustrie, LeipzigGoogle Scholar
  19. 19.
    Horwitz W (1989) J Assoc Off Anal Chem 72:145Google Scholar
  20. 20.
    Wank U (1990) INTERLAB - Ein Auswerte - und Beratungssystem für Ringversuche Diplomarbeit, Friedrich-Schiller-Universität JenaGoogle Scholar
  21. 21.
    Danzer K (1989) Fresenius Z Anal Chem 335:869CrossRefGoogle Scholar
  22. 22.
    Sachs L (1984) Angewandte Statistik, Springer Verlag, Berlin - Heidelberg - New York - TokyoGoogle Scholar
  23. 23.
    Walczak B, Morin-Allory L, Chretien JR, Lafosse M, Dreux M (1986) Chem Intell Lab Sys 1:79CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1991

Authors and Affiliations

  • D. Wienke
    • 1
  • U. Wank
    • 1
  • M. Wagner
    • 1
  • K. Danzer
    • 1
  1. 1.Institute of Inorganic and Analytical Chemistry, Analytical DivisionFriedrich-Schiller-University JenaJenaGermany

Personalised recommendations