统计学习方法第二章——感知机

感知机的实现
参照hankcs统计学习方法的读书笔记中可视化部分（感知机），简单实现了感知机。
关于范数，随机梯度下降这些原理不是很懂，只是根据书中的算法步骤用python实现而已。

#pytho 3.5
#perceptron version 1.0
import numpy as np
from matplotlib import pyplot as plt
from matplotlib import animation
import copy
import pdb

#input: train set T={(x1,y1),...,(xn,yn)},xi in Rn, yi in {-1,1};
def get_input(file = None):# teturn a input set
    # the default data
    return {((1,1),-1),((4,3),1),((3,3),1)}

#input: w,b; model f(x)=sign(w.x+b).
def show_output(result):
    print(result)

#sign
def sign(x):
    if x<0:
        return -1
    else:
        return 1

#the main part
def perceptron(t):  
    history = []
    t_checked = set()
    #give the init value w0,b0
    w=(0,0)
    b=0
    #get the input (xi,yi)
    while(len(t)>0):
        (x,y) = t.pop()
        #if yi(w.xi+b) <= 0,than w=w+n*yi*xi, the default n=1
        r = np.dot(w,x)+b
        print("x:",x,"y:",y,"r:",r)
        if y*r <= 0:
            w = np.add(w,np.dot(y,x))
            b = b+y
            print("update:",w,b)
            t = t.union(t_checked)
            t_checked.clear()
            t.add((x,y))
            history.append([w,b])
        else:
            t_checked.add((x,y))
    result=(w,b)
    show_output(result)
    return history

if __name__ == "__main__":
    training_set = get_input()

#   pdb.set_trace()  #break point
    history = perceptron(copy.copy(training_set))

    # first set up the figure, the axis, and the plot element we want to animate
    fig = plt.figure()
    ax = plt.axes(xlim=(0, 2), ylim=(-2, 2))
    line, = ax.plot([], [], 'g', lw=2)
    label = ax.text([], [], '')

    # initialization function: plot the background of each frame
    def init():
        line.set_data([], [])
        x, y, x_, y_ = [], [], [], []
#       pdb.set_trace()  #break point
        for p in training_set:
            if p[1] > 0:
                x.append(p[0][0])
                y.append(p[0][1])
            else:
                x_.append(p[0][0])
                y_.append(p[0][1])

        plt.plot(x, y, 'bo', x_, y_, 'rx')
        plt.axis([-6, 6, -6, 6])
        plt.grid(True)
        plt.xlabel('x')
        plt.ylabel('y')
        plt.title('Perceptron Algorithm (www.hankcs.com)')
        return line, label


    # animation function.  this is called sequentially
    def animate(i):
        global history, ax, line, label

        w = history[i][0]
        b = history[i][1]
        if w[1] == 0: return line, label
        x1 = -7
        y1 = -(b + w[0] * x1) / w[1]
        x2 = 7
        y2 = -(b + w[0] * x2) / w[1]
        line.set_data([x1, x2], [y1, y2])
        x1 = 0
        y1 = -(b + w[0] * x1) / w[1]
        label.set_text(history[i])
        label.set_position([x1, y1])
        return line, label

    # call the animator.  blit=true means only re-draw the parts that have changed.
    print(history)

#    pdb.set_trace()    #break point

    anim = animation.FuncAnimation(fig, animate, init_func=init, frames=len(history), interval=1000, repeat=True,repeat_delay = 5*1000 ,blit=True)
    plt.show()

输出结果如下：