Model Parameter Sensitivity Analysis Results from Random Forest Ensemble

Statistical analysis results for primary model-related parameters. The dataset includes Feature Importance entries from a 100-tree Random Forest ensemble with over 15,000 samples per network topology and deterministic analytical results from four other methods. It was authored by Chung-Yuan Huang and last updated on 2026-05-21.

Use Cases

• Comparing feature importance metrics between Random Forest and deterministic analytical methods based on the described methods.
• Analyzing the relationship between Parameter Importance (Pearson correlations) and Standardized Regression coefficients based on the orthogonal predictor design.
• Studying the effect of adoption thresholds (avg-of-thresholds) on model outcomes based on the sensitivity sweep from 10 to 70.

Strengths

• Feature Importance estimates have a reported standard error ≤0.001 for all entries.
• Analysis is based on a large sample size (N > 15,000 per network topology).
• Input parameters were varied independently, making correlations approximate zero and simplifying coefficient interpretation.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is very small (9.5 KB), indicating limited scope.

Provenance

Source

figshare

Collection Method

Results from a computational sensitivity analysis, likely from simulation.

Freshness

Last updated 2026-05-21 08:02:10; freshness should be verified.