Release Notes Version 5 5

Version 5.5 of PLS_Toolbox and Solo was released in November, 2009.

For general product information, see PLS_Toolbox Product Page. For information on Solo, see Solo Product Page.

NEW FEATURES

Correlation Spectroscopy GUI added with unique pure component decomposition features to identify correlated spectral components.
Multivariate Curve Resolution by Alternating Least Squares (MCR-ALS) default algorithm is 15-25 times faster than in previous versions.
Direct Scores PLS (DSPLS) algorithm available in PLS (dspls).
Automatic correction of regression model bias and slope based on test data (correctbias). NOTE: PLS_Toolbox only - not yet available in Solo
Bias and R^2 information available for all steps of model building (calibration, cross-validation, validation)
Improved automatic component selection in PCA models.
Additional tests for poor choices of preprocessing with certain model types (e.g. CLS).
User information stored with each model (as history information).

Selectivity ratio (variable selection information) added as calculated statistic for PLS PCR and MLR models (sratio).

KNN Score Distance metric available in scores plots of PLS, PCR, and PCA (knnscoredistance).
Improved handling of class modifications.

Kaiser Optical Systems HoloReact method importer to create preprocessing from band-area methods (hrmethodreadr).
GLSW/EPO: add "meancenter" option to better handle orthogonalization to specific clutter matrix, plus improved help (glsw).
Improved memory and speed performance for various preprocessing methods.
Custom SpectraQuant Filtering algorithm for background correction (available only through Hamilton Sundstrand)

Experiment File Importer - reads file containing filenames and reference measurement data and imports the corresponding files and data into Analysis GUI. Automatically assigns the data to X and Y and to calibration and validation blocks (experimentreadr).
Hamilton Sundstrand PIONIR PDF file importing added (pdfreadr).
Horiba JY file importer performance and support for large files improved (hjyreadr).
SPC file importer reads file comments (spcreadr).
Support for combining multiple DataSets. Selecting multiple items in Workspace Browser provides access to new "combine" menu option.
XY file importing performance improved (xyreadr).
XML importing/exporting support and performance improved (parsexml / encodexml).
Improved support for manual parsing of CSV files (xclreadr).
Export Data to Analect Spectral Files (ASF) (writeasf).
Export Data to simple CSV file format for non-DataSet objects (writecsv).
Export regression Models to CPSA32/PC80 .tbl format (available only through Hamilton Sundstrand).

Plot Controls: Full access to all items in a DataSet (no longer limited by number of rows or columns)
BoxPlot: new box plot function (PLS_Toolbox only) to allow better coexistance with Statistics Toolbox (R) from The Mathworks (boxplot)
Toolbars: Toolbars are now present on most interfaces and plots to make standard actions and interaction between GUIs easier.
TrendTool: Improved interactivity and functionality.
Model Cache: Improved performance and support for new information in models and DataSets.
InfoBox: New File/Save menu option.
DataSet Object: Add ability to log user comments to history.
DataSet Object: Add uniqueID field and support.

boxplot - Box plot showing various statistical properties of a data matrix.

correctbias - Automatically adjust regression model for bias and slope errors.

corrspecgui - Interactive GUI to perform correlation spectroscopy.

dspls - Partial Least Squares computational engine using Direct Scores algorithm.

experimentreadr - Importer for automatic importing and alignment of X and Y blocks.

fasternnls - Fast non-negative least squares with selective constraints.

hrmethodreadr - Convert a Kaiser HoloReact band integration methods into a preprocessing structure.

knnscoredistance - Calculate the average distance to the k-Nearest Neighbors in score space.

pdfreadr - Importer for Hamilton Sundstrand PIONIR PDF files.

sratio - Calculates selectivity ratio for a given regression model.