Gradient Descent for Linear Regression.py

from matplotlib.pyplot import (legend, plot, scatter, show)
from numpy import (array, mean, sqrt, square)

data_set = array([
    [32, 31], [53, 68], [60, 62], [47, 71], [59, 87],
    [55, 78], [52, 79], [39, 59], [48, 75], [52, 71],
    [45, 55], [54, 82], [44, 62], [58, 75], [56, 81],
    [48, 60], [44, 82], [60, 97], [45, 48], [38, 56],
    [66, 83], [65, 11], [47, 57], [41, 51], [51, 75]
])

x_list = data_set[:, 0]
y_list = data_set[:, 1]

count = len(data_set)
intercept = 0
slope = 0


def line_equation(x):
    return slope * x + intercept


learning_rate = 0.0001
iteration_cycles = 1000


for _ in range(iteration_cycles):
    new_y = list(map(line_equation, x_list))
    y_difference = y_list - new_y

    intercept_gradient = sum(-2 / count * y_difference)
    slope_gradient = sum(-2 / count * y_difference * x_list)

    intercept -= learning_rate * intercept_gradient
    slope -= learning_rate * slope_gradient

predicted_y = list(map(line_equation, x_list))

rmse = sqrt(mean(square(y_list - predicted_y)))

print("intercept =", intercept)
print("slope =", slope)
print("rmse =", rmse)

test_x = [44, 32, 63]
test_y = list(map(line_equation, test_x))

scatter(x_list, y_list, label='original')
scatter(test_x, test_y, label='prediction')
plot(x_list, predicted_y, label='line of best fit')
legend()
show()