# 独立成分分析

(Difference between revisions)
 Revision as of 04:34, 8 April 2013 (view source)Kandeng (Talk | contribs)← Older edit Revision as of 04:36, 8 April 2013 (view source)Kandeng (Talk | contribs) Newer edit → Line 73: Line 73: ==中英文对照== ==中英文对照== - :Independent Component Analysis 独立成分分析 + :独立成分分析  Independent Component Analysis - :sparse coding 稀疏编码算法 + :稀疏编码算法 Sparse coding - :over-complete basis 超完备基 + :超完备基      Over-complete basis - :orthonormal basis 标准正交基 + :标准正交基   Orthonormal basis - :sparsity penalty 稀疏惩罚项 + :稀疏惩罚项   Sparsity penalty - :gradient descent 梯度下降法 + :梯度下降法   Gradient descent - :whitened 白化 + :白化         Whitened - :under-complete basis 不完备基 + :不完备基      Under-complete basis - :line-search algorithm 线搜索算法 + :线搜索算法    Line-search algorithm - :topographic cost term 拓扑代价项 + :拓扑代价项    Topographic cost term

## 概述

$J(W) = \lVert Wx \rVert_1$

$\begin{array}{rcl} {\rm minimize} & \lVert Wx \rVert_1 \\ {\rm s.t.} & WW^T = I \\ \end{array}$

## 标准正交ICA

$\begin{array}{rcl} {\rm minimize} & \lVert Wx \rVert_1 \\ {\rm s.t.} & WW^T = I \\ \end{array}$

1. $W \leftarrow W - \alpha \nabla_W \lVert Wx \rVert_1$
2. $W \leftarrow \operatorname{proj}_U W$, 其中U是满足WWT = I的矩阵空间

## 中文译者

Language : English