---
id: "breast-w"
slug: "breast-w"
name: "Breast W"
domain: "Healthcare & Life Sciences"
taskType: "binary_classification"
numRows: 699
numColumns: 10
targetColumn: "diagnosis_class"
primaryMetric: "1 - AUC Error"
primaryScore: "0.0052"
parquetUrl: "https://huggingface.co/datasets/carlahq/demo-tabular-benchmarks/resolve/main/data/breast-w/breast-w.parquet"
spreadsheetUrl: "https://docs.google.com/spreadsheets/d/1NuEqimpAqOjRHz3_Qr4fM8A2puuMuvXUgIxuIl9BunQ/edit?usp=sharing"
canonicalUrl: "https://carlahq.eu/datasets/breast-w"
---

# Dataset: Breast W (breast-w)

> Clinical cytology benchmark containing 699 patient fine needle aspirate (FNA) biopsy records for binary classification of breast tumors into benign or malignant diagnoses based on nuclear morphological characteristics.

## Dataset Overview

| Attribute | Value |
| :--- | :--- |
| **Task Type** | `binary_classification` |
| **Business Domain** | Healthcare & Life Sciences |
| **Target Column** | `diagnosis_class` |
| **Rows (Instances)** | 699 |
| **Features (Columns)** | 10 |
| **Primary Evaluation Metric** | 1 - AUC Error |
| **Best Benchmark Score** | **0.0052** |

## Business Decision Context
**Context:** Early, accurate diagnosis of breast neoplasms from fine needle aspiration (FNA) biopsies is critical in clinical pathology workflows to differentiate non-cancerous lesions from invasive carcinomas without requiring invasive surgical biopsies.
**Objective:** Predict whether a breast mass biopsy sample is benign or malignant using digitized microscopic nuclear morphometric features graded on an ordinal 1-10 clinical scale.
**Decision Impact:** Reduces turnaround time for biopsy triaging and eliminates unnecessary surgical interventions for benign cases. False negatives carry life-threatening diagnostic delays and elevated mortality risks, while false positives result in severe patient psychological distress and costly unnecessary surgical procedures.

## Feature Schema
| Column Name | Data Type | Role | Distinct Values | Missing Values (%) | Description |
| :--- | :--- | :--- | :--- | :--- | :--- |
| `diagnosis_class` | `categorical` | target | 2 | 0 | Final histological and cytological diagnostic determination: benign non-cancerous tumor or malignant carcinoma. |
| `clump_thickness` | `numeric` | feature | 10 | 0 | Grading of whether cells are grouped in mono- or multi-layers; higher values indicate multilayered cellular grouping typical of malignancy (graded 1 to 10). |
| `cell_size_uniformity` | `numeric` | feature | 10 | 0 | Assessment of consistency in cell size across the biopsy sample; high variation indicates anaplasia and malignant progression (graded 1 to 10). |
| `cell_shape_uniformity` | `numeric` | feature | 10 | 0 | Evaluation of nuclear shape consistency and regular cellular geometry versus pleomorphic irregular shapes (graded 1 to 10). |
| `marginal_adhesion` | `numeric` | feature | 10 | 0 | Degree of cohesion between neighboring cells; loss of cell adhesion is a hallmark of malignant invasive potential (graded 1 to 10). |
| `single_epithelial_cell_size` | `numeric` | feature | 10 | 0 | Degree of cellular enlargement in single epithelial cells; significantly enlarged cells correlate with malignancy (graded 1 to 10). |
| `bare_nuclei` | `numeric` | feature | 10 | 16 | Presence of cell nuclei unencapsulated by cytoplasm; elevated bare nuclei counts are strongly associated with benign vs malignant status (graded 1 to 10). |
| `bland_chromatin` | `numeric` | feature | 10 | 0 | Nuclear chromatin texture uniformity; fine/bland texture indicates benignity while coarse, clotted chromatin indicates malignancy (graded 1 to 10). |
| `normal_nucleoli` | `numeric` | feature | 10 | 0 | Prominence and hypertrophy of nucleoli within cell nuclei; large, prominent nucleoli suggest active malignancy (graded 1 to 10). |
| `mitoses_count` | `numeric` | feature | 9 | 0 | Rate and frequency of cell division (mitotic figures) observed under high-power microscopy (graded 1 to 10). |

## Empirical Benchmark Leaderboard
| Rank | Algorithm / Model | Model Family | Runtime | Accuracy | ROC AUC | F1-Score |
| :--- | :--- | :--- | :--- | :--- | :--- | :--- |
| 1 | **AttributeSelectedClassifier PrincipalComponents Ranker J48** | Decision Tree | Python | **97.14%** | 0.9718 | 0.9716 |
| 2 | **A1DE** | Machine Learning Model | Python | **97.14%** | 0.9882 | 0.9715 |
| 3 | **AttributeSelectedClassifier A1DE** | Machine Learning Model | Python | **97.14%** | 0.9882 | 0.9715 |
| 4 | **EXAONE Tabular** | Tabular Foundation Model | PyTorch/Python | **97.00%** | 0.9946 | 0.9572 |
| 5 | **BayesNet K2** | Naive Bayes | Python | **97.00%** | 0.9890 | 0.9701 |
| 6 | **RandomForest** | Random Forest | Python | **97.00%** | 0.9912 | 0.9700 |
| 7 | **AttributeSelectedClassifier RandomForest** | Random Forest | Python | **97.00%** | 0.9917 | 0.9700 |
| 8 | **AttributeSelectedClassifier BayesNet** | Naive Bayes | Python | **97.00%** | 0.9890 | 0.9701 |
| 9 | **BayesNet** | Naive Bayes | Python | **97.00%** | 0.9890 | 0.9701 |
| 10 | **Google TabFM v1.0** | Tabular Foundation Model | PyTorch/Python | **96.97%** | 0.9945 | 0.9567 |
| 11 | **nanotabicl Vanilla** | Tabular Foundation Model | PyTorch/Python | **96.92%** | 0.9948 | 0.9560 |
| 12 | **Streaming nanotabicl** | Tabular Foundation Model | PyTorch/Python | **96.91%** | 0.9948 | 0.9558 |
| 13 | **Full TabICLv2 (PyTorch Reference)** | Tabular Foundation Model | PyTorch/Python | **96.91%** | 0.9948 | 0.9558 |
| 14 | **Carla Engine (TabICLv2 WebGPU)** | Tabular Foundation Model | Browser | **96.90%** | 0.9947 | 0.9556 |
| 15 | **RandomRules** | Machine Learning Model | Python | **96.85%** | 0.9904 | 0.9686 |

### Repeated Cross-Validation Fold Scores (TabICLv2 Reference)
| Fold | Train Samples | Test Samples | Metric Score | Latency |
| :--- | :--- | :--- | :--- | :--- |
| Fold 0 | 466 | 233 | 0.0017 | 671ms |
| Fold 1 | 466 | 233 | 0.0053 | 656ms |
| Fold 2 | 466 | 233 | 0.0066 | 698ms |
| Fold 0 | 466 | 233 | 0.0089 | 669ms |
| Fold 1 | 466 | 233 | 0.0052 | 668ms |
| Fold 2 | 466 | 233 | 0.0038 | 649ms |
| Fold 0 | 466 | 233 | 0.0088 | 683ms |
| Fold 1 | 466 | 233 | 0.0042 | 653ms |
| Fold 2 | 466 | 233 | 0.0041 | 695ms |
| Fold 0 | 466 | 233 | 0.0067 | 664ms |
| Fold 1 | 466 | 233 | 0.0040 | 668ms |
| Fold 2 | 466 | 233 | 0.0035 | 656ms |
| Fold 0 | 466 | 233 | 0.0004 | 710ms |
| Fold 1 | 466 | 233 | 0.0045 | 663ms |
| Fold 2 | 466 | 233 | 0.0118 | 679ms |
| Fold 0 | 466 | 233 | 0.0024 | 661ms |
| Fold 1 | 466 | 233 | 0.0109 | 678ms |
| Fold 2 | 466 | 233 | 0.0028 | 654ms |
| Fold 0 | 466 | 233 | 0.0003 | 670ms |
| Fold 1 | 466 | 233 | 0.0080 | 662ms |
| Fold 2 | 466 | 233 | 0.0060 | 676ms |
| Fold 0 | 466 | 233 | 0.0050 | 673ms |
| Fold 1 | 466 | 233 | 0.0049 | 696ms |
| Fold 2 | 466 | 233 | 0.0055 | 699ms |
| Fold 0 | 466 | 233 | 0.0046 | 665ms |
| Fold 1 | 466 | 233 | 0.0087 | 691ms |
| Fold 2 | 466 | 233 | 0.0011 | 707ms |
| Fold 0 | 466 | 233 | 0.0071 | 683ms |
| Fold 1 | 466 | 233 | 0.0019 | 683ms |
| Fold 2 | 466 | 233 | 0.0046 | 660ms |
| Fold 0 | 466 | 233 | 0.0111 | 683ms |
| Fold 1 | 466 | 233 | 0.0046 | 665ms |
| Fold 2 | 466 | 233 | 0.0016 | 689ms |
| Fold 0 | 466 | 233 | 0.0034 | 656ms |
| Fold 1 | 466 | 233 | 0.0047 | 674ms |
| Fold 2 | 466 | 233 | 0.0054 | 667ms |
| Fold 0 | 466 | 233 | 0.0031 | 709ms |
| Fold 1 | 466 | 233 | 0.0067 | 653ms |
| Fold 2 | 466 | 233 | 0.0074 | 683ms |
| Fold 0 | 466 | 233 | 0.0040 | 653ms |
| Fold 1 | 466 | 233 | 0.0020 | 676ms |
| Fold 2 | 466 | 233 | 0.0101 | 655ms |
| Fold 0 | 466 | 233 | 0.0083 | 703ms |
| Fold 1 | 466 | 233 | 0.0009 | 663ms |
| Fold 2 | 466 | 233 | 0.0061 | 675ms |
| Fold 0 | 466 | 233 | 0.0048 | 653ms |
| Fold 1 | 466 | 233 | 0.0046 | 660ms |
| Fold 2 | 466 | 233 | 0.0076 | 651ms |
| Fold 0 | 466 | 233 | 0.0049 | 678ms |
| Fold 1 | 466 | 233 | 0.0029 | 652ms |
| Fold 2 | 466 | 233 | 0.0051 | 661ms |
| Fold 0 | 466 | 233 | 0.0038 | 658ms |
| Fold 1 | 466 | 233 | 0.0083 | 681ms |
| Fold 2 | 466 | 233 | 0.0029 | 667ms |
| Fold 0 | 466 | 233 | 0.0014 | 662ms |
| Fold 1 | 466 | 233 | 0.0046 | 660ms |
| Fold 2 | 466 | 233 | 0.0119 | 668ms |
| Fold 0 | 466 | 233 | 0.0020 | 683ms |
| Fold 1 | 466 | 233 | 0.0060 | 670ms |
| Fold 2 | 466 | 233 | 0.0081 | 667ms |

## Provenance & Downloadable Artifacts
- **Dataset Parquet File:** [Direct Parquet Download](https://huggingface.co/datasets/carlahq/demo-tabular-benchmarks/resolve/main/data/breast-w/breast-w.parquet)
- **Google Sheets Interactive Demo:** [Open in Google Sheets](https://docs.google.com/spreadsheets/d/1NuEqimpAqOjRHz3_Qr4fM8A2puuMuvXUgIxuIl9BunQ/edit?usp=sharing)
- **OpenML Source:** [OpenML Dataset](https://www.openml.org/d/15)
- **Academic Paper / Citation:** [Read Research Paper](https://doi.org/10.1073/pnas.87.23.9193)
- **Canonical HTML Page:** [https://carlahq.eu/datasets/breast-w](https://carlahq.eu/datasets/breast-w)