Hepatitis Diagnosis Benchmark Data

BNG(hepatitis) is a benchmark dataset for binary classification tasks, likely related to hepatitis diagnosis or patient outcomes. The dataset originates from the OpenML platform, a repository for machine learning data and experiments. Specific details on its size, creator, and creation date are not provided in the available metadata.

Use Cases

• Predict a binary target variable, such as disease presence or patient survival, using patient features like age, sex, and bilirubin levels.
• Benchmark the performance of logistic regression, decision trees, or support vector machines on a medical classification problem.
• Analyze feature importance for the target label to identify key clinical indicators from the dataset's attributes.
• Test data preprocessing techniques for handling potential missing values common in clinical datasets like this one.

Strengths

• Serves as a standardized benchmark for comparing machine learning algorithm performance on a binary classification task.
• Hosted on OpenML, ensuring version control and reproducibility for research.

Limitations

• The exact number of rows and columns is unknown, preventing assessment of dataset scale.
• Lack of column descriptions and sample data obscures the specific features and target variable available.

Provenance

Source

OpenML