logo_2Cy.gif
Home About us Media Research Consultancy Training Site map Contact

Home » Links and downloads » Programs

  • Author: K. Faber
    Purpose: randomization test for PLS component selection
    Description: Matlab functions (zipped = 3 kB) + example data sets (zipped = 20,198 kB) used for:

    • S. Wiklund, D. Nilsson, L. Eriksson, M. Sjöström, S. Wold and K. Faber
      A randomization test for PLS component selection
      Journal of Chemometrics, 21 (2007) 427-439

    Info
  • Author: A. Liwo
    Purpose: estimation of trilinear CANDECOMP/PARAFAC model
    Description: FORTRAN 77 routines (zipped = 615 kB) that use a particular approximation to initialize the least-squares solution by a combination of Marquardt's and Newton's methods. Moreover, an estimate of the uncertainty in the parameter estimates is reported. For further details, see:

    • A. Liwo, P. Skurski, S. Oldziej, L. Lankiewicz, J. Malicka and W. Wiczk
      A new approach to the resolution of the excitation-emission spectra of multicomponent systems
      Computers and Chemistry, 21 (1997) 89-96

    Info
  • Author: A. Olivieri
    Purpose: first-order multivariate calibration
    Description: Matlab toolbox + example data sets + manual (zipped = 1,860 kB) for first-order multivariate calibration (November 2008).
    Info
  • Author: A. Olivieri
    Purpose: second-order multivariate calibration
    Description: Matlab toolbox + example data sets + manual (zipped = 894 kB) for second-order multivariate calibration (November 2008). Includes trilinear decomposition methods (PARAFAC, SWATLD) and residual bilinearization methods (PLS/RBL, BLLS/RBL)
    Info
  • Author: A. Olivieri
    Purpose: third-order multivariate calibration
    Description: MATLAB set of codes, including an example and manual (zipped = 480 kB) for third-order multivariate calibration (January 2011).
    Info
  • Author: A. Olivieri
    Purpose: residual bilinearization
    Description: File (zipped = 114 kB) containing MATLAB routines for performing several residual bilinearization techniques. Includes example data and instructions
    Info
  • Author: A. Pomerantsev
    Purpose: nonlinear modelling
    Description: FITTER, an excell add-in for nonlinear modelling with sophisticated confidence interval estimation. For further details, see:

    • E.V. Bystritskaya, A.L. Pomerantsev and O.Ye. Rodionova
      Non-linear regression analysis: new approach to traditional implementations
      Journal of Chemometrics, 14 (2000) 667-692

    Info
  • Author: R. Wehrens
    Purpose: first-order multivariate calibration
    Description: routines in R for multivariate regression by PLS and PCR, to be downloaded from the LAC site (» Research » Software). As a special feature, the uncertainty in the root mean squared error of crossvalidation (RMSECV) is estimated. Crossvalidation is a popular method for determining the number of factors to include in a calibration model. The uncertainty estimate for RMSECV enables one to select factors on the basis of the significance of a decrease in RMSECV
    Info