Qconcalc and Release Notes Model Exporter Version 3 2: Difference between pages

From Eigenvector Research Documentation Wiki
(Difference between pages)
Jump to navigation Jump to search
imported>Chuck
 
imported>Jeremy
(Created page with "Version 3.2 of Model_Exporter was released in June, 2015. For general product information, see Model_Exporter_User_Guide. ==NEW FEATURES / FIXES== * Add support for non-...")
 
Line 1: Line 1:
===Purpose===
Version 3.2 of Model_Exporter was released in June, 2015. For general product information, see [[Model_Exporter_User_Guide]].


Calculate Q residuals contributions for predictions on a model.
==NEW FEATURES / FIXES==


===Synopsis===
* Add support for non-negative least squares CLS models IF outputting to m-file format (requires function be enabled as sub-functions are added to accomplish the NNLS calculation)
* Add support for approximate nearest neighbor distance via distancemetric option. This is an approximation of the NN distance, but is NOT exact.
* Add support for application of most preprocessing methods to MATRICES
** Allows use of PLS, PCR, PCA, CLS, and PLSDA on x-block of MATRICES (not just vectors) with basically all supported preprocessing methods
* Add basic support for arithmetic on X-block (only - undo on y-block not supported)
* Change "prob" to "probs" in SVMDA so it matches PLSDA outputs
* Fix bug where placeholder variables were not removed from data if SavGol preprocessing was the one and only preprocessing method
* Fix bug which would lead to indexing error in SavGol preprocessing if variables at the END of the spectrum are excluded


:qcon = qconcalc(newx,model)
===Model Interpreter===
:qcon = qconcalc(model);  %requires that model contains residuals


===Description===
* Allow input of matrix instead of just vectors (some methods support applying to matrix)
 
* Force numeric conversion to be done expecting period as decimal separator (resolves problems interpreting models on systems set to other numeric formats like in France)
This function calculates the Q contributions corresponding to the application of either the original calibration data, or an external set of test data, to a PLS or PCA model.
* Move math steps into new MEMath object (simplifies code - exposes mathematical operations to caller)
 
If the model '''model''' was created using the "blockdetails = 'all'" option in [[pls]] or [[pca]] (or whatever function was used to create the model), then '''newx''' can be omitted to retrieve the Q contributions for the calibration data used to build the model. Note, however, that this option is not the default option for these functions, so it is unlikely that this call will work unless you have specifically created the model with the appropriate call.
 
====Inputs====
 
* '''newx''' = matrix containing external test data
* '''model''' = structure for the PLS or PCA model to which the data is to be applied
 
====Outputs====
 
* '''qcon''' = matrix Q contributions for all variables used in '''model''', for all calibration objects, or objects represented in '''newx'''
 
===See Also===
 
[[datahat]], [[pca]], [[pcr]], [[pls]], [[tconcalc]]

Revision as of 15:49, 15 June 2015

Version 3.2 of Model_Exporter was released in June, 2015. For general product information, see Model_Exporter_User_Guide.

NEW FEATURES / FIXES

  • Add support for non-negative least squares CLS models IF outputting to m-file format (requires function be enabled as sub-functions are added to accomplish the NNLS calculation)
  • Add support for approximate nearest neighbor distance via distancemetric option. This is an approximation of the NN distance, but is NOT exact.
  • Add support for application of most preprocessing methods to MATRICES
    • Allows use of PLS, PCR, PCA, CLS, and PLSDA on x-block of MATRICES (not just vectors) with basically all supported preprocessing methods
  • Add basic support for arithmetic on X-block (only - undo on y-block not supported)
  • Change "prob" to "probs" in SVMDA so it matches PLSDA outputs
  • Fix bug where placeholder variables were not removed from data if SavGol preprocessing was the one and only preprocessing method
  • Fix bug which would lead to indexing error in SavGol preprocessing if variables at the END of the spectrum are excluded


Model Interpreter

  • Allow input of matrix instead of just vectors (some methods support applying to matrix)
  • Force numeric conversion to be done expecting period as decimal separator (resolves problems interpreting models on systems set to other numeric formats like in France)
  • Move math steps into new MEMath object (simplifies code - exposes mathematical operations to caller)
Retrieved from "https://www.wiki.eigenvector.com/index.php?title=Qconcalc&oldid=9248"