final_file

final_file

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np

data = pd.read_csv('data.csv')
# print(data)

def cost_function(m, b, points):
    the_total_error = 0
    sum_error = 0
    for i in range(len(points)):
        x = points.iloc[i].weight
        y = points.iloc[i].height
        the_total_error += (y - (m * x + b)) ** 2
        sum_error += the_total_error
    print(sum_error)
    return sum_error

def gradient_descent(m_now, b_now, points, L):
    the_slope_of_thecost_depending_on_m = 0
    the_slope_of_thecost_depending_on_b = 0

    n = len(points)

    for i in range(n):
        x = points.iloc[i].weight
        y = points.iloc[i].height
        the_slope_of_thecost_depending_on_m += - (2/n) * x * (y - (m_now * x + b_now))
        the_slope_of_thecost_depending_on_b += - (2/n) * (y - (m_now * x + b_now))

    m = m_now - L * the_slope_of_thecost_depending_on_m
    b = b_now - the_slope_of_thecost_depending_on_b * L
    return m, b

m = 0
b = 0
Learning_rate = 0.001
times_to_iterate = 10000

#Here we calculate the cost, but it's not needed becuase the gradient descent do that basically for us.
# for i in range(-2, 2):
#    c = cost_function(i, i, data)
# print("The Error")
# print(c)

for i in range(times_to_iterate):
    m, b = gradient_descent(m, b, data, Learning_rate)
print(f"m AND b: {m, b}")
plt.scatter(data['weight'], data['height'])
plt.plot(list(range(0, 5)), [m * x + b for x in range(0 , 5)], color="red")
plt.show()