perceptron.py

import numpy as np
np.random.seed(42)

def stepFunction(t):
    if t >= 0:
        return 1
    return 0

def prediction(X, W, b):
    return stepFunction((np.matmul(X,W)+b)[0])

def perceptronStep(X, y, W, b, learn_rate = 0.01):
    # For all the data points in the dataset
    for i, x in enumerate(X):
        y_cap = prediction(x, W, b)
        
        # If Red point misclassified into Positive region WX + b >= 0
        if y_cap - y[i] == 1:
            W[0] -= learn_rate * x[0]
            W[1] -= learn_rate * x[1]
            b -= learn_rate

        # If Blue point misclassified into Negative region WX + b < 0
        elif y_cap - y[i] == -1:
            W[0] += learn_rate * x[0]
            W[1] += learn_rate * x[1]
            b += learn_rate
    return W, b
  
# Randomly initialise W & b
W = np.array(np.random.rand(2,1))
b = np.random.rand(1)[0] + x_max

# Run the perceptronStep for number of epochs
for i in range(num_epochs):
    W, b = perceptronStep(X, y, W, b, learn_rate)