machine_learn/linear/train.py

#!/usr/bin/python
# -*- coding: UTF-8 -*-
import sys
reload(sys)
sys.setdefaultencoding('utf-8')

import numpy as np
from sklearn.linear_model import LinearRegression
from sklearn import metrics
from draw import draw_util


def curce_data(x,y,y_pred):
	x=x.tolist()
	y=y.tolist()
	y_pred=y_pred.tolist()
	results=zip(x,y,y_pred)
	results=["{},{},{}".format(s[0][0],s[1][0],s[2][0]) for s in results ]
	return results


def read_data(path):
	with open(path) as f :
		lines=f.readlines()
	lines=[eval(line.strip()) for line in lines]
	X,y=zip(*lines)
	X=np.array(X)
	y=np.array(y)
	return X,y


def test():
	X_train,y_train=read_data("train_data")
	X_test,y_test=read_data("test_data")

	#一个对象，它代表的线性回归模型，它的成员变量，就已经有了w，b. 刚生成w和b的时候 是随机的
	model = LinearRegression()
	#一调用这个函数，就会不停地找合适的w和b 直到误差最小
	model.fit(X_train, y_train)
	#打印W
	print model.coef_
	#打印b
	print model.intercept_
	#模型已经训练完毕,用模型看下在训练集的表现
	y_pred_train = model.predict(X_train)
	#sklearn 求解训练集的mse
	# y_train 在训练集上 真实的y值
	# y_pred_train 通过模型预测出来的y值
	#计算  (y_train-y_pred_train)^2/n
	train_mse = metrics.mean_squared_error(y_train, y_pred_train)
	print "训练集MSE:".decode('utf-8'), train_mse

	#看下在测试集上的效果
	y_pred_test = model.predict(X_test)
	test_mse = metrics.mean_squared_error(y_test, y_pred_test)
	print "测试集MSE:".decode('utf-8'),test_mse

	train_curve = curce_data(X_train,y_train,y_pred_train)
	test_curve = curce_data(X_test,y_test,y_pred_test)
	print "推广mse差".decode('utf-8'), test_mse-train_mse

	'''
	with open("train_curve.csv","w") as f :
		f.writelines("\n".join(train_curve))
	with open("test_curve.csv","w") as f :
		f.writelines("\n".join(test_curve))
	'''


def draw_line():
	x_train, y_train = read_data("train_data")
	print(x_train.tolist())
	print(y_train.tolist())
	draw_util.drawScatter(x_train.tolist(), y_train.tolist())


if __name__ == '__main__':
	draw_line()
	test()