Linear_Regression.py

#Importing Modules 
import pandas as pd
import matplotlib.pyplot as plt

data = pd.read_csv('simplelinearregression.csv')

def loss_function(m , b , points):
    total_error = 0
    for i in range(len(points)):
        x = points.iloc[i].Age
        y = points.iloc[i].Premium
        total_error += (y -(m * x + b)) ** 2
    total_error / float(len(points))
    
    #Gradient Descent implementation
def gradient_descent(m_now , b_now , points , L):
    m_gradient = 0
    b_gradient = 0

    n = len(points)

    for i in range(n):
        x = points.iloc[i].Age
        y = points.iloc[i].Premium

        m_gradient += -(2/n) * x * (y-(m_now * x + b_now))
        m_gradient += -(2 / n) * (y - (m_now * x + b_now))

    m  = m_now - m_gradient * L
    b = b_now - b_gradient * L
    return m , b

m = 0
b = 0
L = 0.001
epochs = 300

for i in range(epochs):
    if i % 50 == 0 :
        print(f"Epoch: {i}")
    m , b = gradient_descent(m , b , data , L )

print(m , b )

plt.scatter(data.Age , data.Premium , color="black")
plt.plot(list(range(20,80)), [m * x + b for x in range(20 , 80)] , color='red')
plt.show()