Model Configuration for a Unified Clinical Disease Prediction Framework

A 5.5 KB Excel file containing model configurations for a proposed deep learning framework evaluated on three benchmark clinical datasets. The framework, authored by Vijay U. Rathod, integrates multiple architectures like MLP, CNN, and FT-Transformer. Experimental results from 2026 show the framework achieved an AUC of 0.8980 for heart disease prediction and 0.8451 for diabetes classification.

Use Cases

• Benchmarking deep learning models for heart disease prediction based on the UCI Heart Disease dataset mentioned in the description
• Developing ensemble strategies for diabetes classification based on the PIMA Indians Diabetes dataset mentioned in the description
• Implementing fold-safe feature selection for clinical datasets based on the described robust preprocessing pipeline
• Comparing model performance across datasets with varying feature dimensionality and sample size based on the framework's modular architecture

Strengths

• Framework evaluated on three established benchmark datasets: UCI Heart Disease (303 samples), PIMA Indians Diabetes (768 samples), and a Parkinson's disease voice dataset (195 recordings)
• Reported competitive predictive performance, including an AUC of 0.8980 for heart disease and 0.8451 for diabetes classification
• Released under a permissive CC-BY-4.0 license, facilitating reuse and modification

Limitations

• The dataset is very small (5.5 KB), containing only model configurations, not the underlying clinical data itself
• Column-level documentation for the configuration file is absent; field semantics must be inferred after download
• Row count for the configuration data is unknown, which may limit suitability assessment

Provenance

Source

figshare, authored by Vijay U. Rathod

Collection Method

Likely generated as part of a research study proposing a unified deep learning framework.

Freshness

Last updated 2026-05-08 17:34:18; freshness should be verified.