Batchfold

From Eigenvector Research Documentation Wiki
Revision as of 15:18, 5 September 2012 by imported>Scott (Created page with "===Purpose=== Transform batch data into dataset for analysis. ===Synopsis=== : bdata = batchfold(method,data,options); : [bdata,model] = batchfold(method,data,option...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Purpose

Transform batch data into dataset for analysis.

Synopsis

bdata = batchfold(method,data,options);
[bdata,model] = batchfold(method,data,options);
bdata = batchfold(data,model);

Description

Based on 'method' type, fold/unfold data into suitable dataset for analysis. Data is separated both by batch (high-level experiments) and also optionally by step number (sub-divisions of batch indicating processing segments or other division of batches). Identification of batch and step for each sample must be in .class field. Assumes incoming data is a two-way matrix consisting of samples by variables.ngs.

Inputs

  • method = Method type from table below.
  • data = Dataset object, 2D samples by variables with all batch and step information in the .class field.

Outputs

  • bdata = DataSet Object suitable for loading into 'analysis' interface for given 'method'.
  • model = Standard model structure containing the batchfold model (See MODELSTRUCT). NOTE: Care must be taken to assure fields designated in the calibration set also exist in test set or application of model will fail.

Options

options = a structure array with the following fields:

  • plots: [ {'none'} | 'final' ] governs plotting of results, and
  • order: positive integer for polynomial order {default = 1}.


  • batch_source : [{'class'}|'label'|'axisscale'] Field name of source for batch info. Use 'variable' if selecting a column of data.
  • batch_set : ['BSPC Batch'] Identifies set to use for identifying sample batches. Either a set name (string) or class set number of set to use.
  • batch_locate : {'index'|{'gap'}|'backstep'} How to use variable or axisscale to define steps.
index - boundry at straight index (1 1 1 2 2 3 3 3).
gap - boundry at gaps in data (1 2 3 4 7 8 9 20 21 22).
backstep - boundry at resets (1 2 3 1 2 3 4 1 2).
NOTE: At this point gap and backstep use the same algorithm.
  • step_source : ['class'] Field name of source for step info. Use 'variable' if selecting a column of data. If empty then no steps is assumed.
  • step_set : ['BSPC Step'] Identifies set to use for identifying sample steps. Either a setname (string) or class set number of set to use.
  • step_selection_classes : [] Step numbers (as defined by the step_set) to include in analysis. Empty implies all values are steps. NOTE: This is not an index into the .class field but the actual numeric class values.
  • batch_align_options : [struct] Options for 'batchalign' function. See batchalign for more information.
  • alignment_batch_class : Numeric class of batch to use as reference for alignment or vector of target.
  • alignment_variable_index : Index of variable (columns) in batch to use for alignment.
  • summary : {} Type of summary statistics to calculate for each variable and step (as a cell array of stings). This is only used for spca and sparafac methods.
mean - Mean
std - Standard Deviation
min - Minimum
max- Maximum
range - Range
slope - Slope
length - Length (of step)
percentile - 10 25 50 75 90 percentile.
  • data_only : [{0} | 1 | 2] Only return data:
0 - Run entire function
1 - Make classes for data.
2 - Make classes and align data.

See Also

batchalign, batchdigester, batchmaturity