更新 function.py

function.py

@@ -63,14 +63,28 @@ def load_data(filepath):
         max_columns = max(max_columns, len(parts))
         data.append(parts)
     
-    # 填充缺失的列
+     # 填充缺失的列
     for i in range(len(data)):
         if len(data[i]) < max_columns:
             data[i].extend(['nan'] * (max_columns - len(data[i])))
 
     # 转换成numpy数组，处理缺失值
-    data_array = np.array(data, dtype=float)
-    data_array[np.isnan(data_array)] = np.nanmean(data_array, axis=0)
+    data_array = []
+    for row in data:
+        cleaned_row = []
+        for value in row:
+            if value == '?':  # 替换无法转换的值
+                cleaned_row.append(np.nan)
+            else:
+                cleaned_row.append(float(value))  # 转换为float
+        data_array.append(cleaned_row)
+
+    data_array = np.array(data_array)
+
+    # 处理缺失值
+    col_mean = np.nanmean(data_array, axis=0)
+    for i in range(data_array.shape[1]):
+        data_array[np.isnan(data_array[:, i]), i] = col_mean[i]
 
     return data_array[:, :-1], data_array[:, -1]