Multiresolution Measures

Multiresolution measures analyze complexity at multiple scales.

Purity

Measures class purity in local neighborhoods.

cm = ComplexityMeasures(X, y)
purity = cm.calculate_purity()

Interpretation: - 1.0: Perfect purity (no mixing) - 0.0: Complete mixing

Neighbourhood Separability

Measures how well neighborhoods separate classes.

ns = cm.calculate_neighbourhood_separability()

Interpretation: - Higher values indicate better separation

MRCA - Multiresolution Complexity Analysis

Analyzes complexity across multiple resolutions.

mrca = cm.calculate_MRCA()

Interpretation: - Captures complexity at different scales

C1 - Entropy of Class Proportions

Measures entropy of class distribution.

c1 = cm.calculate_C1()

Interpretation: - Higher values indicate more balanced classes

C2 - Imbalance Ratio

Measures the degree of class imbalance.

c2 = cm.calculate_C2()

Interpretation: - 1.0: Perfectly balanced - Higher values indicate more imbalance

Example: Analyze All Multiresolution Measures

cm = ComplexityMeasures(X, y)
multi = cm.get_all_complexity_measures(measures='multiresolution')

for measure, value in multi.items():
    print(f"{measure}: {value:.4f}")