Done with hybridizing both EASE and CBF model

app.py

@@ -31,11 +31,12 @@ def games_recomm(preferences_id):
             'is_recommended': pref_value
         })
 
-        res = ease_model(pred_df=pred_df, k=10)
-        # res = cbf_model(pred_df=pred_df, k=10)
-    
-    # st.write(res[['predicted_items', 'predicted_score']])
-    st.write(res)
+        res_ease = ease_model(pred_df=pred_df, k=100)
+        res_cbf = cbf_model(pred_df=pred_df, k=100)
+        res = combine_hybrid_result(res_ease, res_cbf)
+        
+    # st.write(res)
+
     if type(res) == ValueError:
         st.error("Recommendation failed. Please select with at least 3 games title.")
         st.session_state
@@ -74,7 +75,6 @@ def spr_sidebar():
 # Home page. One of the page in Main Header
 def home_page():
     st.title("Steam Recommendation System")
-    # st.session_state
 
     if "user_preferences" in st.session_state:
         ids_list = ids_to_titles(st.session_state["user_preferences"])
@@ -106,7 +106,7 @@ def result_page():
         st.error('Please input preferences titles and run "Get recommendation"')
     else:
         st.success(f'Top {len(st.session_state["rs"])}')
-        st.session_state
+        # st.session_state
 
         user_res = generate_res_gamebox(ids=st.session_state['rs'])
 

contentBased_model.py

@@ -103,12 +103,19 @@ class KnnCBF:
         pred_df['predicted_items'] = _output_preds
         pred_df['predicted_score'] = _score_preds
 
-        # escaped = [
-        #     ele for i_list in pred_df['predicted_items'].values for ele in i_list
-        # ]
+        escaped_id = [
+            ele for i_list in pred_df['predicted_items'].values for ele in i_list
+        ]
+        escaped_score = [
+            score for s_list in pred_df['predicted_score'].values for score in s_list
+        ]
+
+        pred_result = pd.DataFrame({
+            'app_id' : escaped_id,
+            'predicted_score' : escaped_score
+        })
         
-        return pred_df
-        # return escaped
+        return pred_result
     
 def cbf_model(pred_df, k=10):
     items = pd.read_csv("data/games_attributes.csv")

easeBased_model.py

@@ -38,6 +38,16 @@ class EASE:
                     "_index"] = dist_labels[col].astype("category").cat.codes
         
         return dist_labels
+    
+    def scaling_score(preds_tensor):
+
+        work_tensor = preds_tensor
+        min_score = torch.min(preds_tensor)
+        max_score = torch.max(preds_tensor)
+
+        scaled =  10 * (work_tensor - min_score) / (max_score - min_score)
+        
+        return scaled
         
     def predict_active(self, pred_df, 
                     weight_mx,
@@ -71,6 +81,10 @@ class EASE:
         )
 
         _preds_act_tensor = user_act_tensor @ weight_mx
+        _preds_act_tensor = 1 * \
+            (_preds_act_tensor - torch.min(_preds_act_tensor)) / \
+            (torch.max(_preds_act_tensor) - torch.min(_preds_act_tensor))
+        
         if remove_owned:
             _preds_act_tensor += -1. * user_act_tensor
 

func.py

@@ -55,8 +55,13 @@ def generate_res_gamebox(ids):
             
             st.divider()
 
-def combine_hybrid_result():
-    pass
+def combine_hybrid_result(res_ease, res_cbf):
+    df = pd.concat([res_ease, res_cbf], axis=0)
+    res_final = df.groupby("app_id").sum().sort_values(["predicted_score"], ascending=False)
+    
+    predictions = res_final.head(10)
+
+    return predictions.index.tolist()
 
     # for title in titles_list:
     #     # url = f"https://store.steampowered.com/app/"