perceptron.py

def preceptron(X, y, lr, epochs):
  m,n = X.shape
  theta = np.zeros((n+1, 1))
  n_miss_list = []
  for epoch in range(epochs):
  n_miss = 0
  for idx, x_i in enumerate(X):
    x_i = np.insert(x_i, 0, 1).reshape(-1, 1)
    y_hat = step_func(np.dot(x_i.T, theta))
    if (np.squeeze(y_hat) - y[idx]) != 0:
      theta += lr * ((y[idx] - y_hat) * x_i) 
      n_miss += 1
   
  n_miss_list.append(n_miss)
 
return theta, n_miss, list