Principal components and representation assumptions
Research context
research method · established · As of 2026-10-10
PCA finds an orthogonal low-rank subspace minimizing squared reconstruction error for centered observations.
Key knowledge
PCA finds an orthogonal low-rank subspace minimizing squared reconstruction error for centered observations. Leading covariance eigenvectors solve this objective; scale choices and outliers change the subspace. Large explained variance need not carry the target signal, and a low-dimensional representation does not establish causal factors. Identifiability can fail when relevant eigenvalues repeat.
Conditions and scope
- Centering, feature units/scaling and rank k are recorded; the criterion uses Euclidean squared error.
- A held-out reconstruction or downstream prediction check is separate from the optimal training projection.
Sources · 1
References
Sources retain their recorded scope. A source link alone does not establish verification of every statement.
Evidence and review
Author checked this English statement, displayed equation where present and stated hypotheses against these source locations: Chapters 2–6 PAC/VC, 11 validation, 13 regularization, 14 stochastic gradient and 23 PCA. The check is bounded to the recorded claims; no complete proof reconstruction, software proof replay or universal literature completeness is claimed.
2026-10-10 · Author source-body/statement/equation/domain check, with graph associations scoped individually. Separate independent content/source/inventory review is required before branch closure.
Entry details
Permanent link to this entry ↗- Record identity
- research-info-representation-pca
- Revision
- research.1
- Structured reference record
- Postgraduate & research reference