Upload app.py

app.py

@@ -7,8 +7,6 @@ from collections import defaultdict
 
 with open('best_svd.pkl', 'rb') as f:
     best_svd = pickle.load(f)
-with open('best_nmf.pkl', 'rb') as f:
-    best_nmf = pickle.load(f)
 with open('model_metadata.pkl', 'rb') as f:
     metadata = pickle.load(f)
 
@@ -16,7 +14,7 @@ movies = metadata['movies_df']
 ratings_filtered = metadata['ratings_filtered_df']
 popular_movies = metadata['popular_movies']
 
-def recommend_movies_gradio(user_id, model_choice, n_recommendations):
+def recommend_movies_gradio(user_id, n_recommendations):
     try:
         user_id = int(user_id)
         n_recommendations = int(n_recommendations)
@@ -37,34 +35,14 @@ def recommend_movies_gradio(user_id, model_choice, n_recommendations):
     all_movies = ratings_filtered['movieId'].unique()
     unseen_movies = [m for m in all_movies if m not in user_ratings]
 
-    if model_choice == "Ensemble (SVD + NMF)":
-        models = [best_svd, best_nmf]
-        ensemble_predictions = defaultdict(list)
-
-        for model in models:
-            for movie_id in unseen_movies:
-                pred = model.predict(user_id, movie_id)
-                ensemble_predictions[movie_id].append(pred.est)
-
-        predictions = []
-        for movie_id, preds in ensemble_predictions.items():
-            predictions.append({
-                'movieId': movie_id,
-                'score': np.mean(preds)
-            })
-    else:
-        if model_choice == "SVD":
-            model = best_svd
-        else:
-            model = best_nmf
-
-        predictions = []
-        for movie_id in unseen_movies:
-            pred = model.predict(user_id, movie_id)
-            predictions.append({
-                'movieId': movie_id,
-                'score': pred.est
-            })
+    # Use SVD model (best performing model - RMSE: 0.9338, Precision@10: 0.7968)
+    predictions = []
+    for movie_id in unseen_movies:
+        pred = best_svd.predict(user_id, movie_id)
+        predictions.append({
+            'movieId': movie_id,
+            'score': pred.est
+        })
 
     predictions_df = pd.DataFrame(predictions)
     top_n = predictions_df.nlargest(n_recommendations, 'score')
@@ -80,27 +58,22 @@ iface = gr.Interface(
     fn=recommend_movies_gradio,
     inputs=[
         gr.Textbox(label="User ID", placeholder="Enter user ID (e.g., 1, 100, 500)"),
-        gr.Dropdown(
-            choices=["Ensemble (SVD + NMF)", "SVD", "NMF"],
-            label="Model Selection",
-            value="Ensemble (SVD + NMF)"
-        ),
         gr.Slider(minimum=5, maximum=50, value=10, step=5, label="Number of Recommendations")
     ],
     outputs=gr.Textbox(label="Recommendations", lines=20),
-    title="<� Movie Recommendation System - MovieLens",
+    title="🎬 Movie Recommendation System - MovieLens",
     description="""
-    Get personalized movie recommendations based on user preferences.
+    Get personalized movie recommendations using SVD (Singular Value Decomposition).
 
-    **Models:**
-    - **Ensemble**: Combines SVD and NMF for robust predictions
-    - **SVD**: Matrix factorization with latent factors
-    - **NMF**: Non-negative matrix factorization
+    **Model Performance:**
+    - **RMSE**: 0.9338 (best prediction accuracy)
+    - **Precision@10**: 0.7968 (79.68% relevant recommendations)
+    - **NDCG@10**: 0.8514 (85.14% ranking quality)
     """,
     examples=[
-        ["1", "Ensemble (SVD + NMF)", 10],
-        ["100", "SVD", 15],
-        ["500", "NMF", 20]
+        ["1", 10],
+        ["100", 15],
+        ["500", 20]
     ]
 )