Cardiac Features For Myocardial Infarction Classification

5,000 patient records contain over 50 extracted features from electrocardiogram (ECG) and photoplethysmogram (PPG) signals for heart attack detection. The dataset is designed for developing machine learning models to classify myocardial infarction. Its origin and last update date are not specified.

Use Cases

• Classify myocardial infarction using 50+ extracted ECG and PPG signal features.
• Train a binary classifier to predict heart attack presence from cardiac waveform metrics.
• Perform feature importance analysis to identify key ECG/PPG indicators for cardiac risk.
• Develop a multi-modal model that fuses ECG and PPG signal characteristics for improved detection.

Strengths

• Contains 5,000 patient records, providing a substantial sample size for model training.
• Offers over 50 distinct cardiac features, enabling detailed feature engineering.

Limitations

• The specific features, column names, and data distributions are unknown, hindering reproducibility.
• Potential class imbalance or label noise is not described, which could affect model performance.
• Geographic and demographic coverage of the 5,000 patients is unspecified, limiting generalizability.

Provenance

Source

Kaggle

Collection Method

Features were likely extracted from ECG and PPG signal recordings, though the exact collection process is unspecified.

Time Range

null

Freshness

null

Geography

null