Parkinson's Disease Stage Prediction Model Based on Six Blood Biomarkers

A dataset used to develop and validate a machine learning model for identifying advanced Parkinson's disease. It contains retrospective data from 536 patients in a discovery cohort and 80 patients in an external validation cohort. The model uses six routine blood biomarkers selected by LASSO and Random Forest algorithms.

Use Cases

• Training a classifier to predict advanced Parkinson's disease stage based on blood biomarker levels.
• Validating the performance of a predictive model on an independent external cohort.
• Analyzing the relationship between specific blood biomarkers (e.g., bilirubin, albumin) and disease progression.

Strengths

• Model demonstrated excellent discriminative ability in the discovery cohort (AUC = 0.873).
• Model maintained robust performance in an independent external validation cohort (AUC = 0.736).
• The dataset includes a total of 616 patient records across two cohorts.

Limitations

• The dataset is very small (17.1 KB), indicating limited scope and feature detail.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

figshare

Collection Method

Retrospective analysis of patient data.

Freshness

Last updated 2026-06-03 05:15:54; freshness should be verified.