import func
if __name__ == "__main__":
print("读取示例数据....")
data1 = pd.read_csv('resources/data1.txt', header=None, names=['x', 'y'])
x = data1['x']
y = data1['y']
m = y.shape[0]
# 给训练的变量加一行1，作为theta0的乘数，方便矩阵运算
x = np.row_stack((np.ones(m), x))
# print(x)
theta = np.zeros(x.shape[0])
print("初始cost值: ", str(func.compute_cost(x, y, theta)))
print("开始进行梯度下降...")
theta = func.gradient_decent(x, y, theta)
print('最终拟合图：')
plt.plot(x[1, :], y, 'r*')
plt.xlabel('x')
plt.ylabel('y')
plt.title('unary linear regression')
plt.grid(True)
plt.plot(x[1, :], x.T.dot(theta))
plt.legend(['First', 'second'], loc=2)
plt.show()
# First task
data = np.matrix(np.loadtxt('ex1data1.txt', delimiter=','))
X = data[:, 0]
y = data[:, 1]
# Second task
plt.plot(X, y, 'g.')
plt.title('Зависимость прибыльности от численности')
plt.xlabel('Численность')
plt.ylabel('Прибыльность')
plt.show()
# Cost function
m = X.shape[0]
X_ones = np.c_[np.ones((m, 1)), X]
theta = np.matrix('[1; 2]')
print(compute_cost(X_ones, y, theta))
# Call method gradient_descent
theta, J_th = gradient_descent(X_ones, y, 0.02, 500)
print(theta)
# Cost change while gradient descent
plt.plot(np.arange(500), J_th, 'k-')
plt.title('Снижение ошибки при градиентном спуске')
plt.xlabel('Итерация')
plt.ylabel('Ошибка')
plt.grid()
plt.show()
# Call method predict
test = np.ones((2, 2))
square = data[:, 0]
room = data[:, 1]
price = data[:, 2]
# нормализация данных
square = normalize(square)
room = normalize(room)
price = normalize(price)
X = data[:, 0:2]
X_ones = np.c_[np.ones((X.shape[0], 1)), X]
y = price
theta = np.matrix('[1; 2; 3]')
# вычисление стоимости
primary_cost = compute_cost(X_ones, y, theta)
print('initial cost -> ' + str(primary_cost))
# градиентный спуск
theta, J_th = gradient_descent(X_ones, y, 0.000000002, 1000)
plt.plot(np.arange(1000), J_th, 'k-')
plt.title('Снижение ошибки при градиентном спуске')
plt.xlabel('Итерация')
plt.ylabel('Ошибка')
plt.grid()
plt.show()
# веса
print('weights:')
print(theta)
import numpy as np
from func import compute_cost, gradient_descent, predict
data = np.matrix(np.loadtxt('ex1data2.txt', delimiter=','))
X = data[:, 0:2]
X = np.c_[np.ones((X.shape[0], 1)), X]
y = data[:, 2]
# Nine task
a = np.linalg.pinv(np.dot(X.T, X))
b = a.dot(X.T)
c = b.dot(y)
print(c)
asd = compute_cost(X, y, c)
print(asd)
test = np.ones((2, 3))
test[0][1] = 272000
test[1][1] = 314000
test[0][2] = 2
test[1][2] = 3
print('prediction ->' + str(predict(test, c)))
data2 = pd.read_csv('resources/data2.txt', header=None)
# print(data2.T)
x = data2.T[0:2].values
x = func.normalize_feature(x)
# print(x)
# 这里为不能用data2.T[2],与dataFrame.ix[n]方式不同，直接使用方括号按行索引取值，时必须指定范围，否则视为按列取
y = data2.T.ix[2].values
print(y)
m = y.shape[0] # 注意，len函数只会计算一列有多少个元素
x = np.row_stack((np.ones(m), x)) # 也可以写作 np.r_[np.ones((1, m)), x]
print(y.shape)
theta = np.zeros(x.shape[0]) # 这里应该根据x的列数进行计算
print("初始cost：", func.compute_cost(x, y, theta))
print("开始梯度下降....")
theta = func.gradient_decent(x, y, theta)
print("最终拟合图3D")
fig = plt.figure()
ax = Axes3D(fig)
X = np.arange(0, 1, 0.01)
Y = np.arange(0, 1, 0.01)
X, Y = np.meshgrid(X, Y) # 将向量变为方阵
def f(x, y):
return theta[0] + theta[1] * x + theta[2] * y
ax.plot_surface(X, Y, f(X, Y), rstride=1, cstride=1)