ML Model Accuracy ANOVA Comparison

import matplotlib.pyplot as plt
import numpy as np
import scipy.stats as stats

np.random.seed(789)

# Cross-validation accuracy scores (100 runs each)
random_forest = np.random.normal(0.892, 0.025, 100)
xgboost = np.random.normal(0.908, 0.022, 100)
neural_net = np.random.normal(0.915, 0.030, 100)
svm = np.random.normal(0.875, 0.028, 100)

F_stat, p_value = stats.f_oneway(random_forest, xgboost, neural_net, svm)

fig, ax = plt.subplots(figsize=(12, 7), facecolor='#0a0a0f')
ax.set_facecolor('#0a0a0f')

colors = ['#6CF527', '#F5B027', '#F5276C', '#4927F5']

parts = ax.violinplot([random_forest, xgboost, neural_net, svm], 
                       positions=[1, 2, 3, 4], showmeans=True, showmedians=True, widths=0.7)

for i, pc in enumerate(parts['bodies']):
    pc.set_facecolor(colors[i])
    pc.set_alpha(0.7)
    pc.set_edgecolor('white')
    pc.set_linewidth(1.5)

parts['cmeans'].set_color('#27D3F5')
parts['cmeans'].set_linewidth(3)
parts['cmedians'].set_color('white')
parts['cmedians'].set_linewidth(1.5)
for partname in ['cbars', 'cmins', 'cmaxes']:
    parts[partname].set_color('#555555')

# Add quartile boxes inside violins
for i, data in enumerate([random_forest, xgboost, neural_net, svm]):
    q1, median, q3 = np.percentile(data, [25, 50, 75])
    ax.vlines(i+1, q1, q3, color='white', linewidth=4, alpha=0.3)
    ax.scatter([i+1], [median], color='white', s=30, zorder=5)

labels = ['Random\nForest', 'XGBoost', 'Neural\nNetwork', 'SVM']
means = [random_forest.mean(), xgboost.mean(), neural_net.mean(), svm.mean()]

# Mean annotations below plot
for i, (mean, color) in enumerate(zip(means, colors)):
    ax.text(i+1, 0.78, f'{mean:.1%}', ha='center', fontsize=11, color=color, fontweight='bold')

# Stats panel - top left, compact
panel_text = f"F={F_stat:.1f}  p={p_value:.2e}  Best: Neural Net ({neural_net.mean():.1%})"
bbox = dict(boxstyle="round,pad=0.4", facecolor='#0d1117', edgecolor='#F5276C', lw=2)
ax.text(0.5, 1.02, panel_text, transform=ax.transAxes, fontsize=10, color='white',
        ha='center', va='bottom', fontfamily='monospace', bbox=bbox)

ax.set_xticks([1, 2, 3, 4])
ax.set_xticklabels(labels, fontsize=10, color='white')
ax.set_ylabel('Classification Accuracy', fontsize=12, color='white', fontweight='500')
ax.set_title('Machine Learning Model Performance\n10-Fold Cross-Validation Results', 
             fontsize=14, color='white', fontweight='bold', pad=25)
ax.set_ylim(0.77, 1.0)

ax.tick_params(colors='#888888')
for spine in ax.spines.values():
    spine.set_color('#333333')
ax.yaxis.grid(True, color='#1a1a2e', linewidth=0.5)
ax.set_axisbelow(True)

plt.tight_layout()
plt.show()