交叉验证knn算法得到最优的参数和准确率不使用sklearn库

交叉验证可以帮助我们在选择超参数时避免过拟合和欠拟合。关于如何实现KNN算法，我这里提供一个简单的实现参考，您可以根据自己的需要进行修改：

import numpy as np

class KNN:
    def __init__(self, k):
        self.k = k
        
    def fit(self, X, y):
        self.X_train = X
        self.y_train = y
        
    def predict(self, X):
        y_pred = []
        for sample in X:
            dists = np.sqrt(np.sum((self.X_train - sample) ** 2, axis=1))
            idxs = np.argsort(dists)[:self.k]
            k_nearest_labels = list(self.y_train[idxs])
            y_pred.append(max(set(k_nearest_labels), key=k_nearest_labels.count))
        return np.array(y_pred)
    
    def accuracy(self, y_true, y_pred):
        return np.sum(y_true == y_pred) / len(y_true)

def cross_validation(X, y, k_fold=5):
    # Split the data into k-fold
    fold_size = len(X) // k_fold
    X_folds = []
    y_folds = []
    for i in range(k_fold):
        X_folds.append(X[i*fold_size:(i+1)*fold_size])
        y_folds.append(y[i*fold_size:(i+1)*fold_size])
        
    # Train and evaluate the model k times
    knn = KNN(k=3) # Change k if necessary
    accuracies = []
    for i in range(k_fold):
        X_train = np.vstack(X_folds[:i] + X_folds[i+1:])
        y_train = np.hstack(y_folds[:i] + y_folds[i+1:])
        X_test = X_folds[i]
        y_test = y_folds[i]
        
        knn.fit(X_train, y_train)
        y_pred = knn.predict(X_test)
        accuracies.append(knn.accuracy(y_test, y_pred))
        
    return np.mean(accuracies), knn.k

上述代码中，我们定义了一个KNN类，用于训练和预测。其中，fit方法接受训练数据和标签，predict方法接受测试数据并返回预测标签，accuracy方法接受真实标签和预测标签并返回准确率。然后，我们定义了cross_validation函数来执行交叉验证。该函数将数据和标签拆分成k个折叠，并在每个折叠上训练和评估模型。最后，该函数返回准确率的平均值和最优的k值。

您可以调用cross_validation函数，以便使用上述代码实现交叉验证knn算法得到最优的参数和准确率。