OCDocker.OCScore.Analysis.Correlation package

This module provides correlation analysis utilities for RMSE and AUC metrics. It computes and visualizes Pearson correlations across methodologies and optionally includes raw and consensus scores.

Usage:

import OCDocker.OCScore.Analysis.Correlation as occorrana

OCDocker.OCScore.Analysis.Correlation.correlation_analysis(results_df, final_metrics, n_trials, error_threshold=1.5, save_path='plots', colour_mapping=None)[source]

Compute and visualize correlation between RMSE and AUC per methodology, including raw and consensus scores.

Parameters:

  • results_df (pd.DataFrame) – DataFrame from parsed Optuna studies with RMSE, AUC and Methodology.

  • final_metrics (pd.DataFrame) – DataFrame with raw scoring and consensus scores.

  • n_trials (int) – Trial count label for output files.

  • error_threshold (float) – Maximum RMSE to include in the filtered subset.

  • save_path (str) – Directory to save output plot.

  • colour_mapping (dict, optional) – Dictionary mapping methodologies to specific colors.

Return type:

None