《统计学习方法》例2.1

demo_2.1.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
# Time-stamp: < 1.py 2016-11-19 14:09:19 >
"""
对应《统计学习方法》李航课本的第2章“感知机”的例题2.1
"""

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]] # 增广矩阵：后面加上+1 -1表示样本的正确输出

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] # check results if divide correctly
	return res

# check if the hyperplane can classify the examples correctly
def check():
	is_need_to_check_next_time = False
	for item in training_set:
		if cal(item) <= 0:
			is_need_to_check_next_time = True
			update(item)
	if not is_need_to_check_next_time:
		print "RESULT: w: " + str(w) + " b: "+ str(b)

	return is_need_to_check_next_time

if __name__=="__main__":
	check_flag = True
	for i in range(1000):# 最大1000次循环
		print i
		check_flag = check()
		if not check_flag:
			break

	# 线性不可分
	if check_flag:
		print "The training_set is not linear separable. "