# Dimensionality reduction

• looking for hidden similarities in data
• based on matrix decomposition

# Example

• Assume that we have 7 Documents with 9 terms.

e.g. Document 1 contains term 6 and term9.

• The document term matrix should be 9×7$\mathbb{R}^{9 \times 7}$, a column represents a document and a raw represents a term.

Remark: we have to normalize our matrix before svd.

• Apply the svd decomposition

M9×7=U9×9Σ9×7VT

• Σ$\Sigma$

• Rank 2 Σ$\Sigma$

UΣ2$U\Sigma_2$ is the 2 rank approximation of the TERM(2 dimension),
Σ2VT$\Sigma_2V^T$ is the 2 rank approximation of the DOCUMENT(2 dimension).

## Question

what do ATA$A^TA$ and AAT$AA^T$ mean if A is a document-term matrix 9×7$\mathbb{R}^{9 \times 7}$?

• ATA7×7$A^TA \in \mathbb{R}^{7 \times 7}$ is the document-document similarity matrix.
• AAT9×9$AA^T \in \mathbb{R}^{9 \times 9}$ is the term-term similarity matrix.

# Latent semantic indexing(LSI, identical to LSA)

• Dimensionality reduction = identification of hidden(latent) concepts
• query matching in latent space

#### 并行计算—使用reduction方法求和

2016-10-25 14:06:46

#### caffe-reduction layer

2016-10-05 20:07:59

#### OpenMp之reduction求和

2014-09-30 09:45:16

#### OpenMP中数据属性相关子句详解(3): reduction子句

2011-11-22 16:12:19

#### Dimension Reduction - feature selection

2014-06-26 11:44:18

#### 规约算法-reduction

2017-05-01 17:46:05

#### scikit-learn：4.4. Unsupervised dimensionality reduction（降维）

2015-07-26 11:14:55

#### 降维 Dimensionality Reduction

2016-06-16 17:13:40

#### TypeError: reduction operation 'argmin' not allowed for this dtype

2018-03-18 12:58:17

#### Barrett Reduction算法求模

2006-05-10 12:56:00

## 不良信息举报

Week3-4Dimensionality reduction