上传新文件

knn.py

+import numpy as np
+from data_proce import one_hot
+
+class Knn:
+    def __init__(self, k=3) -> None:
+        self.k = k
+    
+    def fit(self, X, y):
+        self.X = X
+        self.y = one_hot(y)
+    
+    def predict_proba(self, X_test):
+        if np.ndim(X_test) == 1:
+            X_test = np.expand_dims(X_test, axis=0)
+        result = []
+        for i in range(len(X_test)):
+            dist = (self.X - X_test[i]) ** 2
+            dist = dist.sum(axis=1)
+            samples_id = np.argsort(dist)[:self.k]
+            result.append(self.y[samples_id].mean(axis=0))
+        return np.array(result)
+    
+    def predict(self, X_test):
+        return self.predict_proba(X_test).argmax(axis=1)
\ No newline at end of file