import os import numpy as np import pandas as pd import matplotlib matplotlib.use("Agg") import matplotlib.pyplot as plt import seaborn as sns from sklearn.metrics import ( accuracy_score, precision_score, recall_score, f1_score, roc_auc_score, roc_curve, confusion_matrix, ) FIGURES_DIR = "results/figures" TABLES_DIR = "results/tables" os.makedirs(FIGURES_DIR, exist_ok=True) os.makedirs(TABLES_DIR, exist_ok=True) PALETTE = { "Logistic Regression": "#4878CF", "Random Forest": "#6ACC65", "SVM": "#D65F5F", "Gradient Boosting": "#B47CC7", "ANN (MLP)": "#C4AD66", } # METRICS def evaluate_all_models(fitted_models, X_test, y_test, disease): rows = [] for name, model in fitted_models.items(): y_pred = model.predict(X_test) y_proba = model.predict_proba(X_test)[:, 1] rows.append({ "Model": name, "Accuracy": round(accuracy_score(y_test, y_pred), 4), "Precision": round(precision_score(y_test, y_pred, zero_division=0), 4), "Recall": round(recall_score(y_test, y_pred, zero_division=0), 4), "F1": round(f1_score(y_test, y_pred, zero_division=0), 4), "AUC-ROC": round(roc_auc_score(y_test, y_proba), 4), }) print(f" {name:22s} | Acc={rows[-1]['Accuracy']:.4f} " f"| P={rows[-1]['Precision']:.4f} | R={rows[-1]['Recall']:.4f} " f"| F1={rows[-1]['F1']:.4f} | AUC={rows[-1]['AUC-ROC']:.4f}") df = pd.DataFrame(rows).set_index("Model") df.to_csv(os.path.join(TABLES_DIR, f"{disease}_results.csv")) return df # PLOTS def plot_roc(fitted_models, X_test, y_test, disease): fig, ax = plt.subplots(figsize=(8, 6)) for name, model in fitted_models.items(): y_proba = model.predict_proba(X_test)[:, 1] fpr, tpr, _ = roc_curve(y_test, y_proba) auc = roc_auc_score(y_test, y_proba) ax.plot(fpr, tpr, label=f"{name} (AUC={auc:.3f})", color=PALETTE.get(name, "#999"), linewidth=2) ax.plot([0, 1], [0, 1], "k--", lw=1) ax.set_xlabel("False Positive Rate", fontsize=12) ax.set_ylabel("True Positive Rate", fontsize=12) ax.set_title(f"ROC Curves — {disease}", fontsize=14, fontweight="bold") ax.legend(loc="lower right", fontsize=9) ax.grid(True, alpha=0.3) plt.tight_layout() path = os.path.join(FIGURES_DIR, f"{disease}_roc.png") fig.savefig(path, dpi=150) plt.close(fig) print(f" [Saved] {path}") def plot_metrics_bar(results_df, disease): metrics = ["Accuracy", "Precision", "Recall", "F1", "AUC-ROC"] df_melt = (results_df[metrics].reset_index() .melt(id_vars="Model", var_name="Metric", value_name="Score")) palette = [PALETTE.get(m, "#888") for m in results_df.index] fig, ax = plt.subplots(figsize=(12, 6)) sns.barplot(data=df_melt, x="Metric", y="Score", hue="Model", palette=palette, ax=ax) ax.set_ylim(0, 1.05) ax.set_title(f"Model Performance — {disease}", fontsize=14, fontweight="bold") ax.set_ylabel("Score", fontsize=12) ax.set_xlabel("") ax.legend(bbox_to_anchor=(1.01, 1), loc="upper left", fontsize=9) ax.grid(True, axis="y", alpha=0.3) plt.tight_layout() path = os.path.join(FIGURES_DIR, f"{disease}_metrics_bar.png") fig.savefig(path, dpi=150) plt.close(fig) print(f" [Saved] {path}") def plot_confusion_matrices(fitted_models, X_test, y_test, disease): n = len(fitted_models) cols = min(3, n) rows = (n + cols - 1) // cols fig, axes = plt.subplots(rows, cols, figsize=(cols * 4.5, rows * 4)) axes = np.array(axes).flatten() for i, (name, model) in enumerate(fitted_models.items()): cm = confusion_matrix(y_test, model.predict(X_test)) sns.heatmap(cm, annot=True, fmt="d", cmap="Blues", ax=axes[i], cbar=False, linewidths=0.5) axes[i].set_title(name, fontsize=11, fontweight="bold") axes[i].set_xlabel("Predicted"); axes[i].set_ylabel("Actual") for j in range(i + 1, len(axes)): axes[j].set_visible(False) fig.suptitle(f"Confusion Matrices — {disease}", fontsize=14, fontweight="bold", y=1.01) plt.tight_layout() path = os.path.join(FIGURES_DIR, f"{disease}_confusion.png") fig.savefig(path, dpi=150, bbox_inches="tight") plt.close(fig) print(f" [Saved] {path}") def plot_feature_importance(model, feats, disease, model_name, top_n=15): if not hasattr(model, "feature_importances_"): print(f" [Skip] {model_name} has no feature_importances_") return imp = model.feature_importances_ idx = np.argsort(imp)[::-1][:top_n] fig, ax = plt.subplots(figsize=(9, 6)) ax.barh(range(len(idx)), imp[idx][::-1], color="#4878CF", edgecolor="white") ax.set_yticks(range(len(idx))) ax.set_yticklabels([feats[i] for i in idx][::-1], fontsize=10) ax.set_xlabel("Importance", fontsize=12) ax.set_title(f"Top {len(idx)} Features ({model_name}) — {disease}", fontsize=13, fontweight="bold") ax.grid(True, axis="x", alpha=0.3) plt.tight_layout() path = os.path.join(FIGURES_DIR, f"{disease}_feature_importance.png") fig.savefig(path, dpi=150) plt.close(fig) print(f" [Saved] {path}") # CROSS-DISEASE COMPARISON def plot_cross_disease_heatmap(all_results): summary = pd.DataFrame({d: r["AUC-ROC"] for d, r in all_results.items()}) fig, ax = plt.subplots(figsize=(8, 5)) sns.heatmap(summary, annot=True, fmt=".3f", cmap="YlOrRd", vmin=0.5, vmax=1.0, linewidths=0.5, cbar_kws={"label": "AUC-ROC"}, ax=ax) ax.set_title("AUC-ROC: All Models × All Diseases", fontsize=13, fontweight="bold") ax.set_xlabel("Disease", fontsize=11) ax.set_ylabel("Model", fontsize=11) plt.tight_layout() path = os.path.join(FIGURES_DIR, "cross_disease_auc_heatmap.png") fig.savefig(path, dpi=150) plt.close(fig) print(f" [Saved] {path}") summary.to_csv(os.path.join(TABLES_DIR, "cross_disease_summary.csv")) return summary def plot_cross_disease_bar(all_results): rows = [] for disease, df in all_results.items(): for model_name, row in df.iterrows(): rows.append({"Disease": disease, "Model": model_name, "AUC-ROC": row["AUC-ROC"]}) combined = pd.DataFrame(rows) diseases = list(all_results.keys()) models = combined["Model"].unique() x = np.arange(len(diseases)) width = 0.15 offsets = np.linspace(-(len(models)-1)*width/2, (len(models)-1)*width/2, len(models)) fig, ax = plt.subplots(figsize=(13, 6)) for i, model in enumerate(models): vals = [combined[(combined["Disease"] == d) & (combined["Model"] == model)]["AUC-ROC"].values for d in diseases] vals = [v[0] if len(v) > 0 else 0 for v in vals] ax.bar(x + offsets[i], vals, width, label=model, color=PALETTE.get(model, "#888"), edgecolor="white") ax.set_xticks(x) ax.set_xticklabels(diseases, fontsize=12) ax.set_ylim(0.5, 1.05) ax.set_ylabel("AUC-ROC", fontsize=12) ax.set_title("AUC-ROC Across All Diseases and Models", fontsize=14, fontweight="bold") ax.legend(bbox_to_anchor=(1.01, 1), loc="upper left", fontsize=9) ax.grid(True, axis="y", alpha=0.3) plt.tight_layout() path = os.path.join(FIGURES_DIR, "cross_disease_auc_bar.png") fig.savefig(path, dpi=150) plt.close(fig) print(f" [Saved] {path}")