逻辑回归和线型回归的比较:
1.都是线型模型
2.线型回归的y是连续型的,而逻辑回归是二分类的
3.他们的参数空间都是一致的,信息都蕴含在了w和b中。
由于逻辑回归的结果是分布在0-1区间,所以在线型回归的基础上做一个映射 f f f,即:
f ( x 11 w 1 + x 12 w 2 + x 13 w 3 + . . . + x 1 j w j + . . . + x 1 m w m + b ) = y ^ 1 f( x_{11} w_1 + x_{12} w_2 + x_{13} w_3 + ... + x_{1j} w_j + ... + x_{1m} w_m + b) = \hat y_1 f(x11w1+x12w2+x13w3+...+x1jwj+...+x1mwm+b)=y^1
f ( x 21 w 1 + x 22 w 2 + x 23 w 3 + . . . + x 2 j w j + . . . + x 2 m w m + b ) = y ^ 2 f( x_{21} w_1 + x_{22} w_2 + x_{23} w_3 + ... + x_{2j} w_j + ... + x_{2m} w_m + b) = \hat y_2 f(x21w1+x22w2+x23w3+...+x2jwj+...+x2mwm+b)=y^2
f ( x 31 w 1 + x 32 w 2 + x 33 w 3 + . . . + x 3 j w j + . . . + x 3 m w m + b ) = y ^ 3 f( x_{31} w_1 + x_{32} w_2 + x_{33} w_3 + ... + x_{3j} w_j + ... + x_{3m} w_m + b) = \hat y_3 f(x31w1+x32w2+x33w3+...+x3jwj+...+x3mwm+b)=y^3
. . . ... ...
f ( x n 1 w 1 + x n 2 w 2 + x n 3 w 3 + . . . + x n j w j + . . . + x n m w m + b ) = y ^ n f( x_{n1} w_1 + x_{n2} w_2 + x_{n3} w_3 + ... + x_{nj} w_j + ... + x_{nm} w_m + b) = \hat y_n f(xn1w1+xn2w2+xn3w3+...+xnjwj+...+xnmwm+b)=y^n
这个映射函数我们一般用sigmoid函数,函数形式如下所示:
f ( x ) = 1 1 + e − x f(x) = \frac{1}{1 + e^{-x}} f(x)=1+e−x1
函数图像:
对这个函数求一阶导数:
f ′ ( x ) = ( 1 1 + e − x ) ′ = − 1 ( 1 + e − x ) 2 ( 1 + e − x ) ′ = 1 1 + e − x e − x 1 + e − x = f ∗ ( 1 − f ) \begin{aligned} f'(x) &= (\frac{1}{1 + e^{-x}})'\\ &=- \frac{1}{{(1+e^{-x}})^2} (1+e^{-x})'\\ &=\frac{1}{1+e^{-x}} \frac{e^{-x}}{1+e^{-x}} \\ &=f*(1-f) \end{aligned} f′(x)=(1+e−x1)′=−(1+e−x)21(1+e−x)′=1+e−x11+e−xe−x=f∗(1−f)
方便起见:
x i = [ x i 1 , x i 2 , x i 3 , . . . , x i m ] x_i = [x_{i1}, x_{i2}, x_{i3}, ..., x_{im}] xi=[xi1,xi2,xi3,...,xim]
w = [ w 1 , w 2 , w 3 , . . . , w m ] T w = {[w_1, w_2, w_3, ..., w_m]}^T w=[w1,w2,w3,...,wm]T
则:
y ^ i = f ( x i w + b ) = 1 1 + e − ( x i w + b ) \hat y_i = f(x_i w + b) = \frac{1}{1 + e^{-(x_i w +b)}} y^i=f(xiw+b)=1+e−(xiw+b)1
当 x i w + b → ∞ x_i w +b \to \infin xiw+b→∞, e − ( x i w + b ) → 0 e^{-(x_i w +b)} \to 0 e−(xiw+b)→0, 1 1 + e − ( x i w + b ) → 1 \frac{1}{1 + e^{-(x_i w +b)}} \to 1 1+e−(xiw+b)1→1
故有:
p ( y i = 1 ∣ x i ; w , b ) = f ( x i w + b ) = 1 1 + e − ( x i w + b ) (1) p(y_i=1|x_i; w,b) = f(x_i w + b) = \frac{1}{1 + e^{-(x_i w +b)}} \tag 1 p(yi=1∣xi;w,b)=f(xiw+b)=1+e−(xiw+b)1(1)
p ( y i = 0 ∣ x i ; w , b ) = 1 − f ( x i w + b ) = e − ( x i w + b ) 1 + e − ( x i w + b ) (2) p(y_i=0|x_i; w,b) = 1-f(x_i w + b) = \frac{e^{-(x_i w +b)}}{1 + e^{-(x_i w +b)}} \tag2 p(yi=0∣xi;w,b)=1−f(xiw+b)=1+e−(xiw+b)e−(xiw+b)(2)
为什么说逻辑回归是线型模型?
决策边界来判断
决策边界 p ( y i = 0 ) = p ( y i = 1 ) p(y_i=0) = p(y_i=1) p(yi=0)=p(yi=1)即:
1 1 + e − ( x i w + b ) = e − ( x i w + b ) 1 + e − ( x i w + b ) \frac{1}{1 + e^{-(x_i w +b)}} =\frac{e^{-(x_i w +b)}}{1 + e^{-(x_i w +b)}} 1+e−(xiw+b)1=1+e−(xiw+b)e−(xiw+b)
1 = e − ( x i w + b ) 1= e^{-(x_i w +b)} 1=e−(xiw+b)
( x i w + b ) = 0 (x_i w +b) = 0 (xiw+b)=0
这个决策边界函数是线型函数,所以逻辑回归模型是线型函数。
将(1)和(2)式可合并为:
p ( y i ∣ x i ; w , b ) = [ f ( x i w + b ) ] y i [ 1 − f ( x i w + b ) ] ( 1 − y i ) p(y_i|x_i; w,b) = [f(x_i w + b)] ^{y_i} [1-f(x_i w + b)]^{(1-y_i)} p(yi∣xi;w,b)=[f(xiw+b)]yi[1−f(xiw+b)](1−yi)
对所有样本求似然函数:
L ( w , b ) = ∏ i = 1 n [ f ( x i w + b ) ] y i [ 1 − f ( x i w + b ) ] ( 1 − y i ) L(w,b) = \prod_{i=1}^{n} [f(x_i w + b)] ^{y_i} [1-f(x_i w + b)]^{(1-y_i)} L(w,b)=i=1∏n[f(xiw+b)]yi[1−f(xiw+b)](1−yi)
对数似然函数
log L ( w , b ) = log ∏ i = 1 n [ f ( x i w + b ) ] y i [ 1 − f ( x i w + b ) ] ( 1 − y i ) \log L(w,b) = \log \prod_{i=1}^{n} [f(x_i w + b)] ^{y_i} [1-f(x_i w + b)]^{(1-y_i)} logL(w,b)=logi=1∏n[f(xiw+b)]yi[1−f(xiw+b)](1−yi)
似然函数求取的是最大值,所以损失函数(代价函数)可以定义为:
J ( w , b ) = − log L ( w , b ) = − ∑ i = 1 n y i log [ f ( x i w + b ) ] + ( 1 − y i ) log [ 1 − f ( x i w + b ) ] J(w,b) = -\log L(w, b) = -\sum_{i=1}^n y_i \log [f(x_i w + b)] + (1-y_i) \log [1-f(x_i w + b)] J(w,b)=−logL(w,b)=−i=1∑nyilog[f(xiw+b)]+(1−yi)log[1−f(xiw+b)]
接下来求模型的参数 w , b w, b w,b:
∂ J ( w , b ) ∂ w j = ∂ ∂ w j { − ∑ i = 1 n y i log [ f ( x i w + b ) ] + ( 1 − y i ) log [ 1 − f ( x i w + b ) ] } = − ∑ i = 1 n { y i 1 f ( x i w + b ) − ( 1 − y i ) 1 1 − f ( x i w + b ) } ∂ f ( x i w + b ) ∂ w j = − ∑ i = 1 n { y i 1 f ( x i w + b ) − ( 1 − y i ) 1 1 − f ( x i w + b ) } f ( x i w + b ) [ 1 − f ( x i w + b ) ] ∂ ( x i w + b ) ∂ w j = − ∑ i = 1 n { y i [ 1 − f ( x i w + b ) ] − ( 1 − y i ) f ( x i w + b ) } ∂ ( x i w + b ) ∂ w j = − ∑ i = 1 n { y i [ 1 − f ( x i w + b ) ] − ( 1 − y i ) f ( x i w + b ) } x i j = − ∑ i = 1 n { y i − f ( x i w + b ) } x i j = ∑ i = 1 n { f ( x i w + b ) − y i } x i j = ∑ i = 1 n { y ^ i − y i } x i j \begin{aligned} \frac {\partial J(w,b)}{\partial w_j} &= \frac {\partial}{\partial w_j} \{-\sum_{i=1}^n y_i \log [f(x_i w + b)] + (1-y_i) \log [1-f(x_i w + b)] \} \\ &= -\sum_{i=1}^n \{y_i \frac {1}{f(x_i w + b)} - (1-y_i) \frac {1}{1-f(x_i w + b)} \} \frac {\partial f(x_i w + b)} {\partial w_j} \\ &= -\sum_{i=1}^n \{y_i \frac {1}{f(x_i w + b)} - (1-y_i) \frac {1}{1-f(x_i w + b)} \} f(x_i w + b) [1-f(x_i w + b)] \frac {\partial (x_i w + b)} {\partial w_j} \\ &= -\sum_{i=1}^n \{y_i [1-f(x_i w + b)] - (1-y_i) f(x_i w + b) \} \frac {\partial (x_i w + b)} {\partial w_j} \\ &= -\sum_{i=1}^n \{y_i [1-f(x_i w + b)] - (1-y_i) f(x_i w + b) \} x_{ij} \\ &= -\sum_{i=1}^n \{y_i - f(x_i w + b) \} x_{ij} \\ &= \sum_{i=1}^n \{f(x_i w + b) - y_i \} x_{ij} \\ &= \sum_{i=1}^n \{\hat y_i - y_i \} x_{ij} \\ \end{aligned} ∂wj∂J(w,b)=∂wj∂{−i=1∑nyilog[f(xiw+b)]+(1−yi)log[1−f(xiw+b)]}=−i=1∑n{yif(xiw+b)1−(1−yi)1−f(xiw+b)1}∂wj∂f(xiw+b)=−i=1∑n{yif(xiw+b)1−(1−yi)1−f(xiw+b)1}f(xiw+b)[1−f(xiw+b)]∂wj∂(xiw+b)=−i=1∑n{yi[1−f(xiw+b)]−(1−yi)f(xiw+b)}∂wj∂(xiw+b)=−i=1∑n{yi[1−f(xiw+b)]−(1−yi)f(xiw+b)}xij=−i=1∑n{yi−f(xiw+b)}xij=i=1∑n{f(xiw+b)−yi}xij=i=1∑n{y^i−yi}xij
同理可得:
∂ J ( w , b ) ∂ b = ∑ i = 1 n f ( x i w + b ) − y i = y ^ i − y i \frac {\partial J(w,b)}{\partial b} = \sum_{i=1}^n f(x_i w + b) - y_i = \hat y_i - y_i ∂b∂J(w,b)=i=1∑nf(xiw+b)−yi=y^i−yi
可以用梯度下降法进行对 w w w和 b b b进行求解:
w j = w j − α ∑ i = 1 n ( f ( x i w + b ) − y i ) x i j w_j = w_j - \alpha \sum_{i=1}^{n} {(f(x_i w + b) - y_i)} x_{ij} wj=wj−αi=1∑n(f(xiw+b)−yi)xij
b = b − α ∑ i = 1 n ( f ( x i w + b ) − y i ) b = b - \alpha \sum_{i=1}^{n} {(f(x_i w + b) - y_i)} b=b−αi=1∑n(f(xiw+b)−yi)
也可以随机梯度下降法(SGD):
for i=1 to n:
w j = w j − α ( f ( x i w + b ) − y i ) x i j w_j = w_j - \alpha {(f(x_i w + b) - y_i)} x_{ij} wj=wj−α(f(xiw+b)−yi)xij
b = b − α ( f ( x i w + b ) − y i ) b = b -\alpha {(f(x_i w + b) - y_i)} b=b−α(f(xiw+b)−yi)
当然也可以用批量随机梯度下降法,即循环多次,每次从样本中随机取出一部分样本迭代参数。上式中 α \alpha α为学习率。
import numpy as np
class MyLogisticRegression(object):
def __init__(self, lr=0.01, n_epoch=50):
"""
逻辑回归模型,通过随机梯度下降法实现
:param lr: 学习率,默认值0.01
:param n_epoch: 训练循环次数,默认值50
"""
self.lr = lr
self.n_epoch = n_epoch
self.params = {'w': None, 'b': 0}
def __init_params(self, m):
"""
初始化参数,正态分布
:param m: 特征维度数,和w的个数相对应
:return: None
"""
self.params['w'] = np.random.randn(m)
def fit(self, X, y):
"""
训练模型,随机梯度下降法训练模型
w_j = w_j - lr * (f(x_i @ w) - y_i) * x_ij
b = b - lr * (f(x_i @ w) - y_i)
:param X: 训练数据X,shape(n, m), n代表训练样本个数,m代表每个样本的特征维度数
:param y: 训练数据y, shape(n)
:return: None
"""
X = np.array(X)
# n 样本数量, m 一个样本的特征维度数量
n, m = X.shape
self.__init_params(m)
for _ in range(self.n_epoch):
for x_i, y_i in zip(X, y):
for j in range(m):
self.params['w'][j] = self.params['w'][j] - self.lr * (sigmoid(np.dot(x_i, self.params['w'].T) + self.params['b']) - y_i) * x_i[j]
self.params['b'] = self.params['b'] - self.lr * (sigmoid(np.dot(x_i, self.params['w'].T)+ self.params['b']) - y_i)
def predict(self, X):
"""
预测模型结果
:param X: 二维数据,shape(n,m),n代表个数,m代表单个样本的维度数
:return: y shape(n), 预测结果
"""
return sigmoid(np.dot(X, self.params['w'].T) + self.params['b'])
def sigmoid(x):
return 1 / (1 + np.exp(-x))
def get_samples(n_ex=100, n_classes=2, n_in=1, seed=0):
# 生成100个样本,为了能够在二维平面上画出图线表示出来,每个样本的特征维度设置为1
from sklearn.datasets.samples_generator import make_blobs
from sklearn.model_selection import train_test_split
X, y = make_blobs(
n_samples=n_ex, centers=n_classes, n_features=n_in, random_state=seed
)
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.3, random_state=seed
)
return X_train, y_train, X_test, y_test
def run_my_model():
from matplotlib import pyplot as plt
my = MyLogisticRegression()
X_train, y_train, X_test, y_test = get_samples()
my.fit(X_train, y_train)
y_pred = my.predict(X_test)
print(y_pred, y_test)
# 画图
fig, ax = plt.subplots()
ax.scatter(X_test, y_test, c='red', label='real')
ax.scatter(X_test, y_pred, label='pred')
ax.legend()
plt.show()