A Distance-Preserving Matrix Sketch

Visualizing very large matrices involves many formidable problems. Various popular solutions to these problems involve sampling, clustering, projection, or feature selection reduce the size and complexity of original task. An important aspect methods is how preserve relative distances between points in higher-dimensional space after reducing rows columns fit a lower dimensional space. This because conclusions based on faulty visual reasoning can be harmful. Judging dissimilar as similar basis visualization lead false conclusions. To ameliorate this bias make visualizations datasets feasible, we introduce two new algorithms that, respectively, select subset rectangular matrix. designed closely possible. We compare our matrix sketch more traditional alternatives variety artificial real datasets. Supplementary materials for article are available online.