Selected Wilcoxon Signed-Rank Test Results for Clinical Prediction AUC Scores

5.5 KB of Wilcoxon signed-rank test results comparing deep learning models on three clinical benchmark datasets. The data, authored by Vijay U. Rathod and last updated on 2026-05-08, supports the evaluation of a unified deep learning framework for disease prediction. Results include AUC scores for models like FT-Transformer and CNN ensembles on heart disease, diabetes, and Parkinson's disease detection tasks.

Use Cases

• Compare the statistical significance of AUC differences between deep learning models based on the reported Wilcoxon test results.
• Benchmark new clinical prediction models against the reported AUC scores for heart disease, diabetes, and Parkinson's disease datasets.
• Analyze the performance of ensemble strategies like FT-Transformer + autoencoder based on the provided AUC metrics.
• Validate the reliability of model evaluation using nested cross-validation as referenced in the study description.

Strengths

• Results are derived from three established benchmark datasets: UCI Heart Disease (303 samples), PIMA Indians Diabetes (768 samples), and Parkinson's disease voice (195 recordings).
• Includes specific AUC values with standard deviations, such as 0.8980 (±0.0483) for heart disease prediction.
• Uses robust statistical methods including fold-safe feature selection and nested cross-validation as described.

Limitations

• Row count and specific column definitions are unknown, limiting detailed assessment of the data structure.
• The dataset is very small at 5.5 KB, indicating limited scope, likely containing only summary test statistics.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

figshare, authored by Vijay U. Rathod.

Collection Method

Likely generated from experimental results of a proposed deep learning framework evaluated on benchmark datasets.

Freshness

Last updated 2026-05-08 17:34:36; freshness should be verified.