感知机-Python实现

如图3所示的训练数据集，其正实例点是（3，3），（3,4），负实例点是（1,1），试用感知机学习算法的原始形式求感知机模型，即求出w和b。这里，

图3

这里我们取初值，取。具体问题解释不写了，求解的方法就是算法1。

Python代码如下：

import os

 

# An example in that book, the training set and parameters' sizes are fixed

training_set = [[(3, 3), 1], [(4, 3), 1], [(1, 1), -1]]

 

w = [0, 0]

b = 0

 

# update parameters using stochastic gradient descent

def update(item):

    global w, b

    w[0] = w[0] + 1 * item[1] * item[0][0]

    w[1] = w[1] + 1 * item[1] * item[0][1]

    b = b + 1 * item[1]

    # print w, b # you can uncomment this line to check the process of stochastic gradient descent

 

# calculate the functional distance between 'item' an the dicision surface

def cal(item):

    global w, b

    res = 0

    for i in range(len(item[0])):

        res += item[0][i] * w[i]

    res += b

    res *= item[1]

    return res

 

# check if the hyperplane can classify the examples correctly

def check():

    flag = False

    for item in training_set:

        if cal(item) <= 0:

            flag = True

            update(item)

    if not flag:

        print "RESULT: w: " + str(w) + " b: "+ str(b)

        os._exit(0)

    flag = False

 

if __name__=="__main__":

    for i in range(1000):

        check()

    print "The training_set is not linear separable. "

运行结果如下：