import numpy as np
import matplotlib.pyplot as plt
# 生成从0到100，一共100个元素的数组
x = np.random.randint(0, 100, size = 100)
# 变更数组元素的类型
x = np.array(x, dtype = float)
x_scale = (x - np.min(x)) / (np.max(x) - np.min(x))# 生成50行，2列的矩阵，元素在0到100之间
X = np.random.randint(0, 100, (50, 2))
# 对每一列数据进行最值归一化
X[:, 0] = (X[:, 0] - np.min(X[:, 0])) / (np.max(X[:, 0]) - np.min(X[:, 0]))
X[:, 1] = (X[:, 1] - np.min(X[:, 1])) / (np.max(X[:, 1]) - np.min(X[:, 1]))# 用matplotlib将X展示出来
plt.scatter(X[:, 0], X[:, 1])
plt.show()

X2 = np.random.randint(0, 100, (50, 2))
X2 = np.array(X2, dtype = float)
X2[:, 0] = (X2[:, 0] - np.mean(X2[:, 0])) / np.std(X2[:, 0])
X2[:, 1] = (X2[:, 1] - np.mean(X2[:, 1])) / np.std(X2[:, 1])# 均值接近0
np.mean(X2[:, 0])
# 结果
6.2172489379008772e-17# 方差接近1
np.std(X2[:, 0])
# 结果
0.99999999999999989plt.scatter(X2[:, 0], X2[:, 1])
plt.show()

# 使用鸢尾花数据集
import numpy as np
from sklearn import datasetsiris = datasets.load_iris()
X = iris.data
y = iris.target# 分割出训练数据集和测试数据集
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2, random_state = 666)# 导入StandardScaler，也就是均值方差归一化的对象
from sklearn.preprocessing import StandardScaler
standardScaler = StandardScaler()
standardScaler.fit(X_train)# 将特征训练数据集和特征测试数据集进行归一化处理
X_train_standard = standardScaler.transform(X_train)
X_test_standard = standardScaler.transform(X_test)# 使用kNN
from sklearn.neighbors import KNeighborsClassifier
knn_clf = KNeighborsClassifier(n_neighbors = 3)
knn_clf.fit(X_train_standard, y_train)
knn_clf.score(X_test_standard, y_test)

import numpy as npclass StandardScaler:def __init__(self):self.mean_ = Noneself.scaler_ = None# 获取训练数据集的平均值和方差def fit(self, X):assert X.ndim == 2, "X 的维度必须为2，既X是一个矩阵"self.mean_ = np.array([np.mean(X[:, i]) for i in range(X.shape[1])])self.scaler_ = np.array([np.std(X[:, i]) for i in range(X.shape[1])])return self# 进行均值方差归一化处理def transform(self, X):assert X.ndim == 2, "X 的维度必须为2，既X是一个矩阵"assert self.mean_ is not None and self.scaler_ is not None, "均值和方差不能为空"assert X.shape[1] == len(self.mean_), "训练数据集矩阵的列数必须等于均值数组的元素个数"assert X.shape[1] == len(self.scaler_), "训练数据集矩阵的列数必须等于方差数组的元素个数"X_transform = np.empty(shape=X.shape, dtype=float)for col in range(X.shape[1]):X_transform[:, col] = (X[:, col] - self.mean_[col]) / self.scaler_[col]return X_transform