Phase transition in PCA with missing data: Reduced signal-to-noise ratio, not sample size!
Niels Bruun IpsenLars Kai HansenPublished in: ICML (2019)
Keyphrases
- missing data
- signal to noise ratio
- sample size
- phase transition
- covariance matrix
- principal component analysis
- constraint satisfaction
- missing values
- model selection
- random sampling
- low rank
- np complete
- incomplete data
- worst case
- progressive sampling
- principal components
- dimensionality reduction
- upper bound
- edge detection
- face recognition
- generalization error
- cellular automata
- feature space
- multiscale
- feature extraction
- support vector machine
- evolutionary algorithm
- high dimensional
- data mining