Heart Disease Classification Performance Metrics for a Deep Learning Framework

Vijay U. Rathod's 2026 study proposes a unified deep learning framework for disease prediction, evaluated on three benchmark datasets. The framework achieved an AUC of 0.8980 (±0.0483) for heart disease prediction using an FT-Transformer and autoencoder ensemble. This 5.5 KB Excel file contains the performance results for the heart disease classification task.

Use Cases

• Benchmarking new disease prediction models against the reported AUC of 0.8980 for heart disease classification.
• Studying the performance of modular, dataset-aware deep learning architectures on clinical data.
• Analyzing the impact of ensemble strategies like FT-Transformer + autoencoder on predictive performance.
• Evaluating methods for handling heterogeneous clinical datasets with varying feature dimensionality and sample sizes.

Strengths

• Performance metrics are derived from a framework using robust preprocessing, fold-safe feature selection, and nested cross-validation.
• Specific AUC score of 0.8980 (±0.0483) for heart disease prediction is provided, enabling direct comparison.
• Dataset is openly licensed under CC-BY-4.0, facilitating reuse and verification.

Limitations

• Row count is unknown, which may limit suitability assessment for certain analyses.
• Column-level documentation is absent; field semantics must be inferred after download.
• The 5.5 KB size indicates the file likely contains only summary performance metrics, not the underlying clinical records.

Provenance

Source

figshare

Collection Method

Results from an experimental deep learning study.

Time Range

Study published in 2026.

Freshness

Last updated 2026-05-08 17:34:29; freshness should be verified.

Geography

Benchmark datasets include the UCI Heart Disease dataset; geographic origin of the specific results is not stated.