免责声明:部分内容参考:感知机算法(Perceptron Learning Algorithm)地址https://www.jianshu.com/p/fbfe531ec233
目录
简述:
感知机是二类分类的线性分类模型,其思想很简单,就是
(1)平面内:在一个二维空间中寻找一条直线将两种不同颜色的点集间隔开;
(2)空间内:例如在三维空间内,感知机模型尝试寻找一个超平面,将所有二元类分隔开。
具体的原理、推到再次暂不推到,对跑出来的图形程序暂时展示。
所需的环境搭建:
- 打开cmd
- 输入指令: pip uninstall numpy
- 输入指令:pip install joblib
- 输入指令:pip install numpy
- 输入指令:pip install scipy
图形与代码:
图形1:
![二维平面内](https://i-blog.csdnimg.cn/blog_migrate/1eb4548ae2068ffab1f86b56fd80f66a.png)
图形1的程序:
import numpy as np
import matplotlib.pyplot as plt
X0 = np.array([[1,0],
[0,1],
[2,0],
[2,2]])
X1 = np.array([[-1,-1],
[-1,0],
[-2,-1],
[0,-2]])
#将样本数据化为增广向量矩阵
ones = -np.ones((X0.shape[0],1))
X0 = np.hstack((ones,X0))
ones = -np.ones((X1.shape[0],1))
X1 = np.hstack((ones,X1))
#对样本进行规范化处理
X = np.vstack((-X0,X1))
plt.grid()
plt.scatter(X0[:,1],X0[:,2],c = 'r',marker='o',s=500)
plt.scatter(X1[:,1],X1[:,2],c = 'g',marker='*',s=500)
W = np.ones((X.shape[1],1))
flag = True
while(flag):
flag = False
for i in range(len(X)):
x = X[i,:].reshape(-1,1)
if np.dot(W.T,x)<=0:
W = W + x
flag = True
p1=[-2.0,2.0]
p2=[(W[0]+2*W[1])/W[2],(W[0]-2*W[1])/W[2]]
plt.plot(p1,p2)
plt.show()
图形2:
![](https://i-blog.csdnimg.cn/blog_migrate/bdb6f5310b17c4e85ee9668e88990da3.png)
图形2的程序:
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
#************sigmoid函数计算*********
def sigmoid(z):
return 1/(1+np.exp(-z))
#***********利用迭代方法求最优的w*******
def mymodel(X,Y,eta,W,iters):
i=1
while (i<=iters):
err = sigmoid(np.dot(X,W))-Y
gradient = np.dot(X.T,err)
W=W-eta*gradient
i=i+1
return W
#***********读取数据*******************
df = pd.read_csv('ex1.csv',header=None)
data = df.values
X = data[:,:-1]
Y = data[:,-1].reshape((-1,1))
#**********构造增广矩阵****************
ones = np.ones((X.shape[0],1))
X = np.hstack((ones,X))
#*********模型初始化******************
W = np.ones((X.shape[1],1))
eta = 0.001
#*********调用模型,求最优解**********
W = mymodel(X,Y,eta,W,500)
#*********绘制正例和负例*************
xcord1 = []; ycord1 = []
xcord2 = []; ycord2 = []
for i in range(X.shape[0]):
if int(Y[i])==1:
xcord1.append(X[i,1])
ycord1.append(X[i,2])
else:
xcord2.append(X[i,1])
ycord2.append(X[i,2])
plt.scatter(xcord1,ycord1,s = 20, c = 'red', marker = 's',alpha=.5)
plt.scatter(xcord2, ycord2, s = 20, c = 'green',alpha=.5)
#*********绘制决策边界************
x_w = np.linspace(-3,3,100)
y_w = (-W[0]-W[1]*x_w)/W[2]
plt.plot(x_w,y_w,c='blue')
plt.show()
图形3
![](https://i-blog.csdnimg.cn/blog_migrate/2891c815a348d04beb24d6c4cfeb47b3.png)
暂无代码(插图来自博客地址https://www.jianshu.com/p/fbfe531ec233)