[머신러닝] 선형회귀

import numpy as np
import matplotlib.pyplot as plt

X = np.array([0.0, 1.0, 2.0])
y = np.array([3.0, 3.5, 4.5])

W = 0
b = 0

lrate = 0.01
epochs = 1000

n = float(len(X))

for i in range(epochs):
  y_pred = W * X + b
  dW = (2/n) * sum(X * (y_pred - y))
  db = (2/n) * sum(y_pred - y)
  W = W - lrate * dW
  b = b - lrate * db

plt.scatter(X, y)

plt.plot([min(X),max(X)], [min(y_pred),max(y_pred)], color='red')
plt.show()