https://github.com/htpusa/spca

Sparse Principal Component Analysis via penalised matrix decomposition
https://github.com/htpusa/spca

dimensionality-reduction principal-component-analysis sparse-principal-component-analysis

Last synced: 20 days ago
JSON representation

Sparse Principal Component Analysis via penalised matrix decomposition

• Host: GitHub
• URL: https://github.com/htpusa/spca
• Owner: htpusa
• Created: 2024-02-01T12:57:10.000Z (11 months ago)
• Default Branch: main
• Last Pushed: 2024-04-12T11:54:57.000Z (8 months ago)
• Last Synced: 2024-04-12T18:46:11.824Z (8 months ago)
• Topics: dimensionality-reduction, principal-component-analysis, sparse-principal-component-analysis
• Language: MATLAB
• Homepage:
• Size: 83 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# spca: Sparse Principal Component Analysis via penalised matrix decomposition

MATLAB implementation of the sparse principal component analysis method proposed in Witten et al. (2009).

## Quick start

```MATLAB

load spca_example

[coeff,score] = spca(X,0.5,'K',2);

```

X is a data matrix with observations in rows, the second argument is a parameter between 0 and 1 controlling coefficient vector sparsity: 0 is maximal sparsity, 1 is no sparsity. The name-value argument 'K' controls how many principal components are calculated.

## Optimising the sparsity parameter

```MATLAB

c = tunespca(X,'K',2);

[coeff,score] = spca(X,c,'K',2);

```

The function `tunespca` finds optimal values for the sparsity parameter using the missing value imputation scheme proposed in Witten et al. (2009). The parameter is optimised sequentially for each successive component, `c(1)`is the optimal value for the first component etc.

For more information about the arguments and behaviour, see

```MATLAB

help spca

```

and the example script `spca_example.m`.

## References

Witten, Daniela M., Robert Tibshirani, and Trevor Hastie. "A penalized matrix decomposition, with applications to sparse principal components and canonical correlation analysis." Biostatistics 10.3 (2009): 515-534.