Statistical T-Test P-Values for PepLM-GNN Comparison

17244 benchmark samples underpin the p-values from statistical t-tests comparing the PepLM-GNN model to other baseline methods. The data reports whether PepLM-GNN's accuracy is statistically significantly higher, based on mean ACC values from five-fold cross-validation. Author Ke Yan published this table on figshare under a CC BY 4.0 license.

Use Cases

• Compare PepLM-GNN's performance to baseline methods by analyzing the reported p-values for statistical significance.
• Validate the claim of PepLM-GNN's superior accuracy using the provided t-test results against conventional prediction methods.
• Assess model generalisation performance differences between PepLM-GNN and existing advanced methods based on the cross-validation p-values.

Strengths

• Derived from a substantial benchmark of 17244 samples (8622 positive and 8622 negative pairs).
• Results are based on a rigorous five-fold cross-validation procedure for mean ACC values.
• Published under a permissive CC BY 4.0 license, allowing for reuse and redistribution.

Limitations

• The dataset is very small at 5.5 KB, containing only summary statistical results, not the underlying raw data.
• Limited to a single performance metric (ACC) and statistical test (t-test), omitting other potential comparisons.
• The specific baseline methods compared are not listed in the provided input, limiting interpretability.

Provenance

Source

figshare

Collection Method

Statistical analysis of model performance from five-fold cross-validation on a benchmark dataset.

Freshness

Last updated on 2026-03-24.