Principal component analysis
GPTKB entity
Statements (51)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
statistical analysis
dimensionality reduction technique |
| gptkbp:abbreviation |
gptkb:PCA
|
| gptkbp:application |
gptkb:data_visualization
gptkb:chemometrics exploratory data analysis finance genomics psychology speech recognition compression image compression pattern recognition feature extraction noise reduction face recognition |
| gptkbp:assumes |
large variances have important structure
linearity orthogonality of components |
| gptkbp:category |
unsupervised learning
multivariate statistics |
| gptkbp:form |
orthogonal transformation
eigenvectors of covariance matrix |
| gptkbp:goal |
identify principal components
maximize variance reduce dimensionality of data |
| https://www.w3.org/2000/01/rdf-schema#label |
Principal component analysis
|
| gptkbp:input |
data matrix
|
| gptkbp:introduced |
gptkb:Karl_Pearson
|
| gptkbp:introducedIn |
1901
|
| gptkbp:limitation |
sensitive to scaling
assumes linear relationships components may not be interpretable |
| gptkbp:output |
scores
principal components loadings |
| gptkbp:relatedTo |
gptkb:factor_analysis
eigenvalue decomposition linear algebra singular value decomposition |
| gptkbp:step |
center data
compute covariance matrix compute eigenvectors and eigenvalues project data onto principal components |
| gptkbp:usedIn |
gptkb:machine_learning
gptkb:signal_processing data analysis statistics image processing |
| gptkbp:bfsParent |
gptkb:Boltzmann_machine
|
| gptkbp:bfsLayer |
5
|