# A function to perform a probabilistic principle component analysis

### Description

Performs a probabilistic principle component analysis using the function 'pca' in the package'pcaMethods'

### Usage

1 |

### Arguments

`Data` |
A (non-empty), numeric matrix of data values |

`nPCs` |
The number of resulting principle component axes. nPCs must be less than or equal to the number of columns in Data. |

`CENTER` |
A logical statement indicating whether data should be centered to mean 0, TRUE, or not, FALSE. |

`SCALE` |
A character string indicating which method should be used to scale the variances. The default setting is 'vector.' |

### Details

In PPCA an Expectation Maximization (EM) algorithm is used to fit a Gaussian latent variable model ( Tippping and Bishop (1999)). A latent variable model seeks to relate an observed vector of data to a lower dimensional vector of latent (or unobserved) variables, an approach similar to a factor analysis. Our implementation is a wrapper around the pcaMethods functions ppca and svdimpute (Stacklies et al. (2007)) and is included mainly for convience. The method used in pca was adapted from Roweis (1997) and a Matlab script developed by Jakob Verbeek.

### Value

Returns an object of class 'pcaRes.' See documentation in the package code pcaMethods

### References

Roweis S (1997). EM algorithms for PCA and sensible PCA. Neural Inf. Proc. Syst., 10, 626 - 632.

Stacklies W, Redestig H, Scholz M, Walther D, Selbig J (2007). pcaMethods - a Bioconductor package providing PCA methods for incomplete data. Bioinformatics, 23, 1164 - 1167.

Tippping M, Bishop C (1999). Probabilistic Principle Componenet Analysis. Journal of the Royal Statistical Society. Series B (Statistical Methodology), 61(3), 611 - 622.

### See Also

`pcaMethods`

, `pca`

### Examples

1 2 3 4 |