Visualizes feature importance across the ensemble using scale-invariant metrics. For LASSO models, shows selection frequency (percent of models where coefficient is nonzero). For GLM models, shows significance frequency (percent of models where p < 0.05).
plot_PI_coefficient_summary(coef_summary, main = NULL, label_angle = 45, ...)Output from summarize_PI_coefficients().
Character. Plot title. NULL for auto-generated title based on model type.
Numeric. Rotation angle (degrees) for x-axis labels. Default: 45. Use 0 for horizontal, 90 for vertical. Useful for adjusting label readability in publication figures.
Additional parameters passed to barplot().
Invisible NULL. Plot is drawn to current graphics device.
Coefficient boxplots were intentionally excluded because raw coefficient magnitudes are not comparable across features with different scales (e.g., CADD ~0-50 vs LINSIGHT 0-1). The bar plot uses scale-invariant importance metrics that allow meaningful cross-feature comparison.
if (FALSE) { # \dontrun{
models <- load_PI_models("path/to/piModels/")
coef_summary <- summarize_PI_coefficients(models)
plot_PI_coefficient_summary(coef_summary)
} # }