Preprocessiterator

From Eigenvector Research Documentation Wiki
Jump to navigation Jump to search

Purpose

Create array of preprocessing combinations for use with modeloptimizer.

Synopsis

pplist = preprocessiterator(inpp);%Shows gui for iterator settings.
pplist = preprocessiterator(inpp,imatrix);%Command line call.

Description

For given input preprocessing structure (inpp), create combinations of preprocessing based on PP methods that can be iterated over using simple min/steps/max values. If iteration matrix (imatrix) is not provided a window will appear allowing user to specify iterations. Some of the methods are discussed in Advanced Preprocessing.

Supported Preprocessing Methods:

  1. Derivative (Savgol)
  2. Normalize
  3. GLS Weighting
  4. EPO Filter
  5. Baseline (Automatic Whittaker Filter)
  6. Detrend
  7. Gap Segment Derivative
  8. Autoscale
  9. Poisson (Sqrt Mean) Scaling


Iterator Matrix (imatrix) example. Cell array n x 9 with following columns:

  1. Relative Index - Relative index of given method.
  2. Preprocess Name - Name of preprocess method.
  3. Parameter Name - Name of .userdata parameter.
  4. Parameter Variable - Name of .userdata field.
  5. Data Type - Allowed values for Min and Max.
  6. Min - First value.
  7. Step - Size of interval of each step.
  8. Max - Last value.
  9. Use Log - Use a log scale to create values.


inpp = preprocess('default','mean center','derivative','normalize', 'mean center','sqmnsc','normalize','log10','whittaker');

imatrix = { 1 'derivative' 'Width' 'width' 'int(1:inf)' 1 1 1 0;
1 'derivative' 'Derivative' 'deriv' 'int(1:inf)' 1 1 1 0;...
1 'derivative' 'Order' 'order' 'int(1:inf)' 1 1 1 0;
2 'Normalize' 'Norm Type' 'normtype' 'int(1:inf)' 1 2 2 0;...
1 'GLS Weighting' 'Alpha' 'a' 'float(0:inf)' 1 1 1 1};

pplist = preprocessiterator(inpp,imatrix)

NOTE: If the original preprocess structure contains 2 Normalize steps, the second Normalize will be iterated over.

See Also

preprocess, preprouser