第一阶段-第8天-NumPy入门

今日目标

1. 安装并熟悉NumPy
2. 理解数组与列表的区别
3. 掌握基本的NumPy运算
4. 理解向量化的好处

学习内容

1. 安装NumPy

pip3 install numpy

验证安装：

import numpy as np
print(np.__version__)

2. 创建数组

import numpy as np

# 从列表创建
a = np.array([1, 2, 3, 4, 5])
print(a)           # [1 2 3 4 5]
print(type(a))     # <class 'numpy.ndarray'>

# 创建特定形状的数组
b = np.zeros(5)    # [0. 0. 0. 0. 0.]
c = np.ones(5)     # [1. 1. 1. 1. 1.]
d = np.arange(5)   # [0 1 2 3 4]
e = np.linspace(0, 10, 5)  # [ 0.  2.5  5.  7.5 10.]

print("zeros:", b)
print("ones:", c)
print("arange:", d)
print("linspace:", e)

3. 数组的基本运算

import numpy as np

a = np.array([1, 2, 3, 4, 5])
b = np.array([10, 20, 30, 40, 50])

# 逐元素运算
print("加法:", a + b)        # [11 22 33 44 55]
print("减法:", a - b)        # [-9 -18 -27 -36 -45]
print("乘法:", a * b)        # [ 10  40  90 160 250]
print("除法:", b / a)        # [10. 10. 10. 10. 10.]

# 标量运算
print("乘以2:", a * 2)       # [ 2  4  6  8 10]
print("平方:", a ** 2)       # [ 1  4  9 16 25]

# 内置函数
print("求和:", np.sum(a))     # 15
print("平均值:", np.mean(a))  # 3.0
print("最大值:", np.max(a))   # 5
print("最小值:", np.min(a))   # 1

4. 多维数组

import numpy as np

# 二维数组（矩阵）
矩阵 = np.array([
    [1, 2, 3],
    [4, 5, 6],
    [7, 8, 9]
])

print("矩阵:\n", 矩阵)
print("形状:", 矩阵.shape)   # (3, 3)
print("行数:", 矩阵.shape[0])  # 3
print("列数:", 矩阵.shape[1])  # 3

# 按列计算
print("每列的和:", np.sum(矩阵, axis=0))  # [12 15 18]
# 按行计算
print("每行的和:", np.sum(矩阵, axis=1))  # [ 6 15 24]

5. 向量化 vs 循环

import numpy as np
import time

n = 1000000

# 方法1：纯Python循环
a_list = list(range(n))
b_list = list(range(n, 2*n))

start = time.time()
结果 = [x + y for x, y in zip(a_list, b_list)]
python_time = time.time() - start

# 方法2：NumPy向量化
a_np = np.arange(n)
b_np = np.arange(n, 2*n)

start = time.time()
结果 = a_np + b_np
numpy_time = time.time() - start

print(f"Python循环: {python_time:.4f}秒")
print(f"NumPy向量化: {numpy_time:.4f}秒")
print(f"加速比: {python_time / numpy_time:.1f}倍")

通常NumPy比纯Python快10-100倍。

今日练习

练习1：用NumPy重新计算MSE

import numpy as np

真实值 = np.array([1, 3, 5, 7, 9])
预测值 = np.array([1.5, 2.8, 5.2, 6.9, 9.5])

# 一行代码计算MSE
mse = np.mean((预测值 - 真实值) ** 2)
print(f"MSE = {mse:.4f}")

练习2：批量计算预测值

import numpy as np

输入 = np.array([1, 2, 3, 4, 5])
w = 2.0
b = 1.0

# 向量化预测
预测值 = w * 输入 + b
print("预测值:", 预测值)
# 输出: [ 3.  5.  7.  9. 11.]

练习3：矩阵乘法

import numpy as np

# 两个矩阵相乘
A = np.array([[1, 2], [3, 4]])
B = np.array([[5, 6], [7, 8]])

C = np.dot(A, B)  # 或者 A @ B
print("A × B =\n", C)
# [[19 22]
#  [43 50]]

验收标准

• 能安装并使用NumPy
• 能创建一维和二维数组
• 能进行基本的数组运算
• 能解释向量化相比循环的优势