EVOLUTION-MANAGER
Edit File: ica.html
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><title>R: Independent Component Analysis</title> <meta http-equiv="Content-Type" content="text/html; charset=utf-8" /> <link rel="stylesheet" type="text/css" href="R.css" /> </head><body> <table width="100%" summary="page for ica {e1071}"><tr><td>ica {e1071}</td><td style="text-align: right;">R Documentation</td></tr></table> <h2>Independent Component Analysis</h2> <h3>Description</h3> <p>This is an R-implementation of the Matlab-Function of Petteri.Pajunen@hut.fi. </p> <p>For a data matrix X independent components are extracted by applying a nonlinear PCA algorithm. The parameter <code>fun</code> determines which nonlinearity is used. <code>fun</code> can either be a function or one of the following strings "negative kurtosis", "positive kurtosis", "4th moment" which can be abbreviated to uniqueness. If <code>fun</code> equals "negative (positive) kurtosis" the function tanh (x-tanh(x)) is used which provides ICA for sources with negative (positive) kurtosis. For <code>fun == "4th moments"</code> the signed square function is used. </p> <h3>Usage</h3> <pre> ica(X, lrate, epochs=100, ncomp=dim(X)[2], fun="negative") </pre> <h3>Arguments</h3> <table summary="R argblock"> <tr valign="top"><td><code>X</code></td> <td> <p>The matrix for which the ICA is to be computed</p> </td></tr> <tr valign="top"><td><code>lrate</code></td> <td> <p>learning rate</p> </td></tr> <tr valign="top"><td><code>epochs</code></td> <td> <p>number of iterations</p> </td></tr> <tr valign="top"><td><code>ncomp</code></td> <td> <p>number of independent components</p> </td></tr> <tr valign="top"><td><code>fun</code></td> <td> <p>function used for the nonlinear computation part</p> </td></tr> </table> <h3>Value</h3> <p>An object of class <code>"ica"</code> which is a list with components </p> <table summary="R valueblock"> <tr valign="top"><td><code>weights</code></td> <td> <p>ICA weight matrix</p> </td></tr> <tr valign="top"><td><code>projection</code></td> <td> <p>Projected data</p> </td></tr> <tr valign="top"><td><code>epochs</code></td> <td> <p>Number of iterations</p> </td></tr> <tr valign="top"><td><code>fun</code></td> <td> <p>Name of the used function</p> </td></tr> <tr valign="top"><td><code>lrate</code></td> <td> <p>Learning rate used</p> </td></tr> <tr valign="top"><td><code>initweights</code></td> <td> <p>Initial weight matrix</p> </td></tr> </table> <h3>Note</h3> <p>Currently, there is no reconstruction from the ICA subspace to the original input space.</p> <h3>Author(s)</h3> <p>Andreas Weingessel</p> <h3>References</h3> <p>Oja et al., “Learning in Nonlinear Constrained Hebbian Networks”, in Proc. ICANN-91, pp. 385–390. </p> <p>Karhunen and Joutsensalo, “Generalizations of Principal Component Analysis, Optimization Problems, and Neural Networks”, Neural Networks, v. 8, no. 4, pp. 549–562, 1995. </p> <hr /><div style="text-align: center;">[Package <em>e1071</em> version 1.7-3 <a href="00Index.html">Index</a>]</div> </body></html>