Principal Component Analysis

A linear dimensionality reduction technique that transforms high-dimensional data into a set of uncorrelated components to maximize variance.

Karl Pearson introduced PCA in 1901 to distill complex datasets into their most informative signals. The process identifies orthogonal axes (principal components) that capture the highest degree of variance within the data. Modern data scientists use it to compress 1000-plus feature sets (like genomic sequences or pixel arrays) into 2 or 3 dimensions for visualization or noise reduction. The algorithm leverages Singular Value Decomposition (SVD) to rank these components by their eigenvalue magnitude. It remains the primary tool for streamlining machine learning models and identifying hidden structures in high-frequency financial data.

https://scikit-learn.org/stable/modules/decomposition.html#pca