---
license: cc-by-4.0
author : Christopher Alberto Hamonangan
---

# Overview

CLB-Patient-Screening is a synthetic benchmark designed for tabular anomaly
detection in healthcare, focusing on multi-biomarker systemic drift and early
signs of chronic metabolic disorders.

## Dataset information

| Property                    | Description                                       |
| --------------------------- | ------------------------------------------------- |
| Dataset name                | CLB-Patient-Screening                             |
| Full name                   | Clinical Laboratory Biomarker Panel               |
| Publisher                   | BhinnekaAI                                        |
| Dataset Creator             | Christopher Alberto Hamonangan                    |
| Domain                      | Healthcare and clinical laboratory screening      |
| Data modality               | Structured numerical tabular data                 |
| Feature count               | 5 numerical features                              |
| Data type                   | Synthetic                                         |
| Task                        | Pathological anomaly detection                    |
| Expected anomaly proportion | Approximately 5.0%                                |
| Label convention            | `0` = healthy patient; `1` = pathological anomaly |

# Dataset Structure

## random-seed

- Directory: random-seed
- Files: 5 files, 5,000 observations each.

## seed42

- Directory: seed42
- Files: 5 files with sizes 100, 1,000, 10,000, 50,000, and 100,000.

# Feature Specification

| Feature              | Description                                 | Data Type  |
| -------------------- | ------------------------------------------- | ---------- |
| `blood_pressure`     | Systolic blood pressure (mmHg)              | Continuous |
| `blood_glucose`      | Fasting blood glucose level (mg/dL)         | Continuous |
| `ldl_cholesterol`    | Low-density lipoprotein cholesterol (mg/dL) | Continuous |
| `white_blood_cells`  | White blood cell count ($10^3/\mu L$)       | Continuous |
| `inflammatory_index` | Composite systemic inflammation score       | Continuous |

# Mathematical Formulation

Normal biomarkers are modeled using a Multivariate Gaussian distribution
reflecting strict physiological homeostasis:

x_normal ~ N(μ, Σ)

where μ = [120, 90, 130, 6.5, 1.8] and Σ represents biological covariance constraints.

# Loading the Dataset

```python
import pandas as pd

df = pd.read_csv("seed42/dataset_10000_rows_seed42.csv")
X = df.iloc[:, :-1].values
y = df['is_pathological'].values
```