Y-Randomization Validation Results for Predictive Models (100 Iterations per Model)

100 randomization iterations were performed for each model–predictor–target combination to assess model robustness. The dataset contains evaluation metrics like R², Q², RMSE, MAE, and p-values computed on randomized target variables. Authored by Mythili V and last updated on 2026-05-21, it is hosted on figshare under a CC-BY-4.0 license.

Use Cases

• Validate predictive model robustness based on performance metrics from randomized data.
• Compare original model performance against chance correlations using the provided R2_fit_random and Q2_LOOCV_random scores.
• Demonstrate statistical reliability of a model in research publications using the p_value_random field.

Strengths

• Contains 100 randomization iterations for each model–predictor–target combination, providing a substantial basis for statistical assessment.
• Includes multiple evaluation metrics (R², Q², RMSE, MAE, correlation, p-value) for a detailed performance comparison.
• Explicitly structured with fields for Iteration, Predictor, Target, Model, and Family, clarifying the experimental design.

Limitations

• Row count is unknown, which may limit suitability assessment for large-scale analyses.
• Column-level documentation is absent; field semantics must be inferred after download from the high-level description.

Provenance

Source

figshare

Collection Method

Results from Y-randomization (response permutation) experiments.

Freshness

Last updated 2026-05-21 14:34:11; freshness should be verified.