更新 function.py

function.py

@@ -56,6 +56,27 @@ def classify(input_vector, dataSet, labels, k=2):
     sorted_class_count = sorted(class_count.items(), key=lambda x: x[1], reverse=True)
     return sorted_class_count[0][0]
 
+def cross_validation(dataSet, labels, k_list=(2, 3, 5, 7, 9)):
+    from sklearn.model_selection import KFold
+    kf = KFold(n_splits=5)
+    best_k = None
+    best_accuracy = 0
+    for k in k_list:
+        accuracies = []
+        for train_index, test_index in kf.split(dataSet):
+            train_data, test_data = dataSet[train_index], dataSet[test_index]
+            train_labels, test_labels = np.array(labels)[train_index], np.array(labels)[test_index]
+            predictions = []
+            for vector in test_data:
+                predicted_label = classify(vector.reshape(1, -1), train_data, train_labels, k)
+                predictions.append(predicted_label)
+            accuracy = sum([1 for p, t in zip(predictions, test_labels) if p == t]) / len(test_labels)
+            accuracies.append(accuracy)
+        avg_accuracy = sum(accuracies) / len(accuracies)
+        if avg_accuracy > best_accuracy:
+            best_accuracy = avg_accuracy
+            best_k = k
+    return best_k
 
 def img_to_vector(filename):
     """
@@ -100,8 +121,9 @@ def load_dataset(directory):
 
 
 if __name__ == "__main__":
-    # 以下代码可用于测试，可自行随意修改
     train_data, train_labels = load_dataset("trainingDigits")
+    best_k = cross_validation(train_data, train_labels)
+    print("Best k value:", best_k)
 
     test_data = (
         "00000000000000110000000000000000"
@@ -141,6 +163,7 @@ if __name__ == "__main__":
         np.array([int(char) for char in test_data]).reshape(1, 1024),
         train_data,
         train_labels,
+        best_k
     )
 
     print(f"预测的标签: {predicted_label}")