批量梯度下降
```
import math
import numpy as np
from matplotlib import pyplot
a=[150,200,250,300,350,400,600]#面积
p=[6450,7450,8450,9450,11450,15450,18450]#价格
ste=0.000000001#步长
theta0=300
theta1=1
m=7
temp1=0
temp2=0
k=0
b1=1
b2=0
while math.fabs(theta0-temp1)>0.0000001 and math.fabs(theta1-temp2)>0.0000001:#当参数ans1与真实值相差小于0.0000001
# while math.fabs(b1-b2)>0.0001:
sum = 0 #并且参数ans2与真实值相差也小于0.0000001时结束迭代
k+=1 #次数统计
for i in range(0, 7):
sum +=theta0 + theta1 * a[i] - p[i]
sum /= m
sum1=0
for i in range(0, 7):
sum1 += (theta0+ theta1 * a[i] - p[i])*a[i]
sum1 /= m
temp2 = theta1
theta0 = theta0 - sum * ste # 生成新的theta0
theta1 = theta1 - sum1* ste # 生成新的theta1
b2 = b1
for i in range(0,7):
b1=( theta0 +theta1* a[i] -p[i])**2
b1/=(2*m)
print(k,theta0,theta1,k) #表示正在运行，输出点东西要好些吧,为了和结果区分，对称一下，绘图的第三方库安装出现问题了，输出数据来小替代一下吧
print(k,theta0,theta1);#输出参数theta0与参数theta1
#打印图形
pyplot.scatter(a, p)
x = np.arange(100, 700, 100)
y = theta1 * x + theta0
pyplot.plot(x, y)
pyplot.xlabel('area')
pyplot.ylabel('price')
pyplot.show()
```
随机梯度下降
```
import math
import random
import numpy as np
from matplotlib import pyplot
a=[150,200,250,300,350,400,600]#面积
p=[6450,7450,8450,9450,11450,15450,18450]#价格
ste=0.000000001#步长
theta0=300
theta1=1
m=7
temp1=0
temp2=0
k=0
b1=1
b2=0
while math.fabs(theta0-temp1)>0.0000001 and math.fabs(theta1-temp2)>0.0000001:#当参数ans1与真实值相差小于0.0000001
# while math.fabs(b1-b2)>0.0001:
#并且参数ans2与真实值相差也小于0.0000001时结束迭代
k+=1 #次数统计
temp2 = theta1
i=random.randint(0,6)
theta0 = theta0 - (theta0 + theta1 * a[i] - p[i]) * ste # 生成新的theta0
theta1 = theta1 - (theta0+ theta1 * a[i] - p[i])*a[i]* ste # 生成新的theta1
b2 = b1
for i in range(0,7):
b1=( theta0 +theta1* a[i] -p[i])**2
b1/=m
print(k,theta0,theta1,k) #表示正在运行，输出点东西要好些吧,为了和结果区分，对称一下，绘图的第三方库安装出现问题了，输出数据来小替代一下吧
print(k,theta0,theta1);#输出参数theta0与参数theta1
#打印图形
pyplot.scatter(a, p)
x = np.arange(100, 700, 100)
y = theta1 * x + theta0
pyplot.plot(x, y)
pyplot.xlabel('area')
pyplot.ylabel('price')
pyplot.show()
```
批量梯度等了几小时，还没跑完，先来一个随机梯度的结果吧，
又过了几个小时emmm，我改了一下步长再交一下粗略结果吧
批量梯度下
```
import math
import numpy as np
#from matplotlib import pyplot
a=[1.1,1.3,1.5,2,2.2,2.9,3,3.2,3.2,3.7,3.9,4,4,4.1,4.5,4.9,5.1,5.3,5.9,6,6.8,7.1,7.9,8.2,8.7,9,9.5,9.6,10.3,10.5]#工作经验
p=[39343,46205,37731,43525,39891,56642,60150,54445,64445,57189,63218,55794,56957,57081,61111,67938,66029,83088,81363,93940,91738,98273,101302,113812,109431,105582,116969,112635,122391,121872]#薪水
ste=0.00001#步长
theta0=30000
theta1=1
m=30
temp1=0
temp2=0
k=0
b1=1
b2=0
while math.fabs(theta0-temp1)>0.0000001 and math.fabs(theta1-temp2)>0.0000001:#当参数ans1与真实值相差小于0.0000001
# while math.fabs(b1-b2)>0.0001:
sum = 0 #并且参数ans2与真实值相差也小于0.0000001时结束迭代
k+=1 #次数统计
for i in range(0, 30):
sum +=theta0 + theta1 * a[i] - p[i]
sum /= m
sum1=0
for i in range(0, 30):
sum1 += (theta0+ theta1 * a[i] - p[i])*a[i]
sum1 /= m
temp2 = theta1
theta0 = theta0 - sum * ste # 生成新的theta0
theta1 = theta1 - sum1* ste # 生成新的theta1
b2 = b1
for i in range(0,30):
b1=( theta0 +theta1* a[i] -p[i])**2
b1/=(2*m)
print(k,theta0,theta1,k) #表示正在运行，输出点东西要好些吧,为了和结果区分，对称一下，绘图的第三方库安装出现问题了，输出数据来小替代一下吧
print(k,theta0,theta1);#输出参数theta0与参数theta1
#打印图形
pyplot.scatter(a, p)
x = np.arange(10000, 70000, 10000)
y = theta1 * x + theta0
pyplot.plot(x, y)
pyplot.xlabel('area')
pyplot.ylabel('price')
pyplot.show()
```
随机梯度下降
```
import math
import random
import numpy as np
from matplotlib import pyplot
a=[1.1,1.3,1.5,2,2.2,2.9,3,3.2,3.2,3.7,3.9,4,4,4.1,4.5,4.9,5.1,5.3,5.9,6,6.8,7.1,7.9,8.2,8.7,9,9.5,9.6,10.3,10.5]#工作经验
p=[39343,46205,37731,43525,39891,56642,60150,54445,64445,57189,63218,55794,56957,57081,61111,67938,66029,83088,81363,93940,91738,98273,101302,113812,109431,105582,116969,112635,122391,121872]#薪水
ste=0.000001#步长
theta0=30000
theta1=1
m=30
temp1=0
temp2=0
k=0
b1=1
b2=0
while math.fabs(theta0-temp1)>0.0000001 and math.fabs(theta1-temp2)>0.0000001:#当参数ans1与真实值相差小于0.0000001
# while math.fabs(b1-b2)>0.0001:
#并且参数ans2与真实值相差也小于0.0000001时结束迭代
k+=1 #次数统计
temp2 = theta1
i=random.randint(0,29)
theta0 = theta0 - (theta0 + theta1 * a[i] - p[i]) * ste # 生成新的theta0
theta1 = theta1 - (theta0+ theta1 * a[i] - p[i])*a[i]* ste # 生成新的theta1
b2 = b1
for i in range(0,29):
b1=( theta0 +theta1* a[i] -p[i])**2
b1/=m
print(k,theta0,theta1,k) #表示正在运行，输出点东西要好些吧,为了和结果区分，对称一下，绘图的第三方库安装出现问题了，输出数据来小替代一下吧
print(k,theta0,theta1);#输出参数theta0与参数theta1
#打印图形
pyplot.scatter(a, p)
x = np.arange(100, 700, 100)
y = theta1 * x + theta0
pyplot.plot(x, y)
pyplot.xlabel('area')
pyplot.ylabel('price')
pyplot.show()
```
多元线性回归
批量梯度下降
```
import numpy as np
from matplotlib import pyplot
a=[[1,2],[2,1],[2,3],[3,5],[1,3],[4,2],[7,3],[4,5],[11,3],[8,7]]#数据集一
p=[7,8,10,14,8,13,20,16,28,26]#数据二
ste=0.00001#步长
theta0=5
theta1=1
theta2=1
m=10
temp1=0
temp2=0
temp3=0
k=0
b1=1
b2=0
while math.fabs(theta0-temp1)>0.0000001 and math.fabs(theta1-temp2)>0.0000001 and math.fabs(theta2-temp3)>0.0000001:#当参数ans1与真实值相差小于0.0000001
# while math.fabs(b1-b2)>0.0001:
sum = 0 #并且参数ans2与真实值相差也小于0.0000001时结束迭代
k+=1 #次数统计
for i in range(0, 10):
sum +=theta0 + theta1 * a[i][0]+theta2*a[i][1] - p[i]
sum /= m
sum1=0
sum2=0
for i in range(0, 10):
sum1 += (theta0+ theta1 * a[i][0] +theta2*a[i][1]- p[i])*a[i][0]
sum2 += (theta0+ theta1 * a[i][0] +theta2*a[i][1]- p[i])*a[i][1]
sum1 /= m
sum2 /= m
temp2 = theta1
temp3 = theta2
theta0 = theta0 - sum * ste # 生成新的theta0
theta1 = theta1 - sum1* ste # 生成新的theta1
theta2 = theta2 - sum2* ste # 生成新的theta2
b2 = b1
for i in range(0,10):
b1=( theta0 +theta1* a[i][0] +theta2*a[i][1] -p[i])**2
b1/=(2*m)
print(k,theta0,theta1,theta2,k) #表示正在运行，输出点东西要好些吧,为了和结果区分，对称一下，绘图的第三方库安装出现问题了，输出数据来小替代一下吧
print(k,theta0,theta1,theta2);#输出参数theta0与参数theta1
#打印图形
pyplot.scatter(a, p)
x = np.arange(100, 700, 100)
y = theta1 * x + theta0
pyplot.plot(x, y)
pyplot.xlabel('area')
pyplot.ylabel('price')
pyplot.show()
```
随机梯度下降
```
import math
import random
import numpy as np
from matplotlib import pyplot
a=[[1,2],[2,1],[2,3],[3,5],[1,3],[4,2],[7,3],[4,5],[11,3],[8,7]]#数据集一
p=[7,8,10,14,8,13,20,16,28,26]#数据二
ste=0.00000001#步长
theta0=5
theta1=1
theta2=1
m=10
temp1=0
temp2=0
temp3=0
k=0
b1=1
b2=0
while math.fabs(theta0-temp1)>0.0000001 and math.fabs(theta1-temp2)>0.0000001 and math.fabs(theta2-temp3)>0.0000001:#当参数ans1与真实值相差小于0.0000001# while math.fabs(b1-b2)>0.0001:
#并且参数ans2与真实值相差也小于0.0000001时结束迭代
k+=1 #次数统计
temp2 = theta1
temp3 = theta2
i=random.randint(0,9)
theta0 = theta0 - (theta0 + theta1 * a[i][0]+theta2*a[i][1] - p[i]) * ste # 生成新的theta0
theta1 = theta1 - (theta0+ theta1 * a[i][0] +theta2*a[i][1]- p[i])*a[i][0]* ste # 生成新的theta1
theta2 = theta2 - (theta0+ theta1 * a[i][0] +theta2*a[i][1]- p[i])*a[i][1]* ste # 生成新的theta2
b2 = b1
for i in range(0,10):
b1=( theta0 +theta1* a[i][0]+theta2*a[i][1] -p[i])**2
b1/=m
print(k,theta0,theta1,theta2,k) #表示正在运行，输出点东西要好些吧,为了和结果区分，对称一下，绘图的第三方库安装出现问题了，输出数据来小替代一下吧
print(k,theta0,theta1,theta2);#输出参数theta0与参数theta1
#打印图形
pyplot.scatter(a, p)
x = np.arange(100, 700, 100)
y = theta1 * x + theta0
pyplot.plot(x, y)
pyplot.xlabel('area')
pyplot.ylabel('price')
pyplot.show()
```
x和y的尺寸没调好暂时这样吧
网友评论