Upload 2 files

app.py

@@ -0,0 +1,126 @@
+# -*- coding: utf-8 -*-
+"""ITI105_SETY_Demo_Final.ipynb
+
+Automatically generated by Colab.
+
+Original file is located at
+    https://colab.research.google.com/drive/1239rYpEr1h2-cpIWEzcYdiaG6QSZCNzu
+"""
+
+import gradio as gr
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+import joblib
+from sklearn.metrics import (
+    accuracy_score, precision_score, recall_score, f1_score,
+    roc_curve, precision_recall_curve, auc,
+    confusion_matrix, ConfusionMatrixDisplay
+)
+
+# Load models (adjust paths as needed)
+base_model = joblib.load("base_model.pkl")
+best_model = joblib.load("best_gb.pkl")
+
+def get_metrics(model, X, y):
+    y_pred = model.predict(X)
+    y_prob = model.predict_proba(X)[:, 1]
+
+    acc = accuracy_score(y, y_pred)
+    prec = precision_score(y, y_pred)
+    rec = recall_score(y, y_pred)
+    f1 = f1_score(y, y_pred)
+
+    fpr, tpr, _ = roc_curve(y, y_prob)
+    precision_vals, recall_vals, _ = precision_recall_curve(y, y_prob)
+
+    return {
+        'y_pred': y_pred,
+        'y_prob': y_prob,
+        'fpr': fpr,
+        'tpr': tpr,
+        'precision_vals': precision_vals,
+        'recall_vals': recall_vals,
+        'metrics': f"""
+        📌 Accuracy: {acc:.4f}
+        📌 Precision: {prec:.4f}
+        📌 Recall: {rec:.4f}
+        📌 F1 Score: {f1:.4f}
+        📌 ROC AUC: {auc(fpr, tpr):.4f}
+        📌 PR AUC: {auc(recall_vals, precision_vals):.4f}
+        """
+    }
+
+
+def plot_confusion_matrices(model1, model2, X, y, labels=["Base Model", "Best Model"]):
+    fig, axes = plt.subplots(1, 2, figsize=(16, 8))
+    #fig, axes = plt.subplots(1, 2, figsize=(50, 25))
+
+    for i, model in enumerate([model1, model2]):
+        y_pred = model.predict(X)
+        cm = confusion_matrix(y, y_pred)
+        disp = ConfusionMatrixDisplay(confusion_matrix=cm)
+        disp.plot(ax=axes[i], cmap='Blues', colorbar=False)
+        # Manually annotate the matrix with larger font
+        for (j, k), val in np.ndenumerate(cm):
+            axes[i].text(k, j, f"{val}", ha='center', va='center', fontsize=28, color='red')
+        axes[i].set_title(f"{labels[i]} Confusion Matrix", fontsize=30)
+        axes[i].tick_params(axis='both', labelsize=36)
+
+    plt.tight_layout()
+    return plt.gcf()
+
+def evaluate(file):
+    df = pd.read_csv(file.name)
+    if 'Status' not in df.columns:
+        return "Error: 'Status' column missing."
+
+    #df = df.tail(5)
+    X = df.drop(columns='Status')
+    y = df['Status']
+
+    base = get_metrics(base_model, X, y)
+    best = get_metrics(best_model, X, y)
+
+    # Combined ROC Curve
+    plt.figure()
+    plt.plot(base['fpr'], base['tpr'], label=f"Base Model (AUC={auc(base['fpr'], base['tpr']):.2f})", linestyle='--')
+    plt.plot(best['fpr'], best['tpr'], label=f"Best Model (AUC={auc(best['fpr'], best['tpr']):.2f})", linestyle='-')
+    plt.plot([0, 1], [0, 1], 'k--', alpha=0.5)
+    plt.xlabel("False Positive Rate")
+    plt.ylabel("True Positive Rate")
+    plt.title("Combined ROC Curve")
+    plt.legend()
+    roc_fig = plt.gcf()
+
+    # Combined PR Curve
+    plt.figure()
+    plt.plot(base['recall_vals'], base['precision_vals'], label=f"Base Model (AUC={auc(base['recall_vals'], base['precision_vals']):.2f})", linestyle='--')
+    plt.plot(best['recall_vals'], best['precision_vals'], label=f"Best Model (AUC={auc(best['recall_vals'], best['precision_vals']):.2f})", linestyle='-')
+    plt.xlabel("Recall")
+    plt.ylabel("Precision")
+    plt.title("Combined Precision-Recall Curve")
+    plt.legend()
+    pr_fig = plt.gcf()
+
+    # Confusion Matrices
+    cm_fig = plot_confusion_matrices(base_model, best_model, X, y)
+
+    combined_metrics = f"📊 Base Model:\n{base['metrics']}\n\n📊 Best Model:\n{best['metrics']}"
+
+    return combined_metrics, roc_fig, pr_fig, cm_fig
+
+demo = gr.Interface(
+    fn=evaluate,
+    inputs=gr.File(label="Upload CSV with 'Status' column"),
+    outputs=[
+        gr.Textbox(label="📊 Performance Comparison"),
+        gr.Plot(label="Combined ROC Curve"),
+        gr.Plot(label="Combined Precision-Recall Curve"),
+        gr.Plot(label="Confusion Matrices")
+    ],
+    title="🔍 Model Comparison Dashboard",
+    description="Upload a CSV file to compare base and best model performance side by side."
+)
+
+demo.launch(debug=True)

requirements.txt

@@ -0,0 +1,5 @@
+gradio
+pandas
+scikit-learn
+matplotlib
+joblib