Chris Beaumont's IDL Library

Download source code

single page | use frames     summary     class     fields     routine details     file attributes

./

pricom__define.pro

Class description for pricom

Routines

result = pricom::project(data [, nterm=nterm] [, coeffs=coeffs])

Projects a given vector onto the vector space defined by the principal compents, and returns the result

result = pricom::get_pc()

Return the principal components.

result = pricom::get_eval()
result = pricom::get_mean()
result = pricom::init(data)

Build a new PCA object.

pricom::cleanup
pricom__define

This class provides a more convenient interface for working with principal component analysis.

Routine details

toppricom::project

result = pricom::project(data [, nterm=nterm] [, coeffs=coeffs])

Projects a given vector onto the vector space defined by the principal compents, and returns the result

Return value

The projection of the input onto the principal components. This is equivalent to sum(coeff[i] * PC_i) where coeff[i] = sum(data * PC_i)

Parameters

data in required

A vector to project

Keywords

nterm in optional

Set to an integer to project onto only the first nterm principal components.

coeffs in optional

Set to a variable to hold the coefficients of the linear combination of principal components. In other words, result = sum( coeff[i] * PC_i )

Author information

History:

March 2010: Written by Chris Beaumont April 2010: Fixed a bug that crashed this function when called with only one data point. cnb April 2010: Number of returned coeffs matches nterm. cnb. August 2010: Truncated intermediate arrays when nterm < ndim. Speeds up execution. cnb. August 10 2010: Further optimization when nterm < ndim. cnb. August 18 2010: Fixed spelling error in definition of get_mean. cnb.

toppricom::get_pc

result = pricom::get_pc()

Return the principal components.

toppricom::get_eval

result = pricom::get_eval()

toppricom::get_mean

result = pricom::get_mean()

toppricom::init

result = pricom::init(data)

Build a new PCA object.

Parameters

data in required

an (ndim) by (nobj) array of training data. Contains (nobj) vectors, each with (ndim) elements. These are the data used to generate the principal components

toppricom::cleanup

pricom::cleanup

toppricom__define

pricom__define

This class provides a more convenient interface for working with principal component analysis. The user supplies a training data set, which is used to generate principal components via the IDL procedure PCOMP. The methods in this class then allow a simple way to project new vectors onto PCA space, obtain the principal components, etc.

Author information

History:

April 2010: Written by Chris Beaumont June 11 2010: Class now uses the cnb_pca function to calculate principal components. cnb.

File attributes

Modifcation date: Wed Aug 18 19:40:51 2010
Lines: 144