startown/wisconsin-diagnostic-breast-cancer

Wisconsin Diagnostic Breast Cancer (WDBC)

This tabular dataset contains 569 samples, each a fine-needle aspirate (FNA) of a breast mass. Each row holds 30 nuclear features computed from a digitized image of the sample, plus a diagnosis label.

On each image, several cell nuclei are outlined and 10 measurements are taken from every nucleus:

1. 1.radius: mean distance from the center to points on the perimeter
2. 2.texture: standard deviation of gray-scale values
3. 3.perimeter: length of the nuclear contour
4. 4.area: number of pixels inside the contour
5. 5.smoothness: local variation in radius lengths
6. 6.compactness: perimeter^2 / area - 1.0
7. 7.concavity: severity of indentations in the contour
8. 8.concave points: number of indentations in the contour
9. 9.symmetry: difference in length between opposite radial lines
10. 10.fractal dimension: "coastline approximation" - 1

These ten values are then summarized across the nuclei in that image as a mean, a standard error, and a worst value (the mean of the three largest). The diagnosis label marks the whole sample as B (benign) or M (malignant).

The 569 samples are split into a training set (80%, 455) and a test set (20%, 114) with a fixed random seed 42. The ID column is dropped. The text labels B and M are converted to numerical targets 0 (benign) and 1 (malignant).

The data is from the UC Irvine Machine Learning Repository. To cite it, use:

@misc{breast_cancer_wisconsin_(diagnostic)_17,
  author       = {Wolberg, William, Mangasarian, Olvi, Street, Nick, and Street, W.},
  title        = {{Breast Cancer Wisconsin (Diagnostic)}},
  year         = {1993},
  howpublished = {UCI Machine Learning Repository},
  note         = {{DOI}: https://doi.org/10.24432/C5DW2B}
}