ParLeS executable software for chemometric analysis of spectroscopic data.
This software version 3.1
A description of ParLeS is published in
the journal Chemometrics & Intelligent Laboratory Systems:
Viscarra Rossel, R.A. 2007. ParLeS: Software for chemometric analysis of
spectroscopic data. Chemometrics and Intelligent Laboratory Systems, (in
press)
doi:10.1016/j.chemolab.2007.06.006
ParLeS v3.1 can be used to: (i.) pre-process data (iii.) perform PCA, (iv.) PLSR1 delete-n-cross validation (v.) PLSR1 modelling, (v.) PLSR1 prediction, (vi) bootstrap aggregation PLSR (bagging-PLSR) and (vii) join multiple spectroscopic files from single directory into a single file. The software also provides options to save all output for external plotting and analyses. Version 3.1 supports interactive outlier detection.
This is what it looks like:
 |
 |
|
Figure 1. Front page and instructions on how to run
software. |
Figure 2. Import spectral X data and y-variables for modelling and prediction/testing. |
 |
 |
| Figure 3. Transform, preprocess and pretreat data using a combination of methods, as required. | Figure 4. Perform principal components analysis (PCA) on X data. Outliers detected by point and click |
 |
 |
| Figure 5. Perform delete-n jackknife cross validation. RMSE, Rsq. Qsq. and AIC statistics provided for selection of optimal number of factors. Outliers detected by point and click. | Figure 6. Model data using PLSR1. Graphics of scores, regression coefficients, loadings, weights and variable importance for projection (VIP). Data identified by point and click |
 |
 |
| Figure 7. Use the PLSR model for predictions. Statistics of predictions shown. Outliers detected by point and click | Figure 8. Use bagging-PLSR for more robust predictions. Bagging-PLSR will also give you a measure of prediction uncertainty. Outliers detected by point and click |
If you
would like to use ParLeS, please send me an email (r.viscarra-rossel@usyd.edu.au)
briefly explaining your application and your current situation (i.e.
student, researcher, working at university, commercial organisation,
etc.).
If ParLeS is used in research or commercially please cite the following reference:
Viscarra Rossel, R.A. 2007. ParLeS: Software for chemometric analysis
of spectroscopic data. Chemometrics and Intelligent Laboratory Systems
(in-press)
doi:10.1016/j.chemolab.2007.06.006
For more up to date citation information you may also visit my
Publications page
Partial Least Squares (PLS) regression (also known as PLSR or PLSR1) is a popular modelling technique in chemometrics, econometrics and in industrial applications. It is a data compression technique that is also commonly used in spectral quantitative analysis. Research in science often involves using variables that are easily (or cheaply) measured to explain or predict the behaviour of response variables that are often much more difficult (or expensive) to acquire. When the factors are few in number and are not significantly redundant (or collinear) and have a well understood relationship to the responses, then multiple linear regression (MLR) can be a useful way to turn data into information. However, if these conditions break down, then MLR will not be efficient or appropriate. PLS is a method used to construct predictive models when factors are many and highly collinear, e.g. in reflectance sprectroscopy. The emphasis of PLSR is on predicting the respose. However, when used interactively with proper graphics and validation, it also allows the user to attain a good causal insight into the underlying relationhisp between the variables.
PLSR is closely related to Principal Component Regression (PCR). However, PLSR is performed in a slightly different manner. Take the case where we want to use spectral reflectance data to model and then estimate the value/concentration of a soil property: instead of first decomposing the spectra into a set of eigenvectors and scores, and regressing them against the soil values as a separate step, PLSR actually uses the soil information during the decomposition process (the decomposition of both the spectral and the soil data into their most common variations is performed simultaneously). PLSR takes advantage of the correlation that exists between the spectra and the soil values. So, the resulting spectral vectors are directly related to the soil values/concentrations.
For more details see references at the bottom of this page or the following web resource www.statsoft.com/textbook/stpls.html.
Can also find lots of information on spectroscopy and chemometric methodologies at the following sites:
Spectroscopy
Europe - Tony Davies Column
Near Infrared Site
Disclaimer
I have taken all care to ensure that ParLeS is operationally sound. However, this program is supplied 'as is' and no warranty is provided or implied. I assume no liability for damages, direct or consequential that may result from the use of ParLeS.
References
Geladi, P. and Kowalski, B.R. 1986. Partial least-squares regression: a tutorial. Analytica Chimica Acta, 185: 1-17.
Martens H., Nęs, T. (1989). Multivariate Calibration. John Wiley & Sons, Chichester.
Viscarra Rossel, R.A. (2007)
Robust modelling of soil diffuse reflectance spectra by 'bagging-'PLSR'.
Journal of Near Infrared Spectroscopy 15: 39-47.
Viscarra Rossel, R.A. (2007) ParLeS: Software for chemometric
analysis of spectroscopic data. Chemometrics and Intelligent Laboratory
Systems (in press)
doi:10.1016/j.chemolab.2007.06.006