Machine Learning Derivatives

ml_derivatives.py

import numpy as np

def predict(X, w): 
    return X * w
    
def loss (X, Y, w): 
    return np.average ((predict(X, w) - Y) ** 2)

def gradient(X, Y, w): 
    w_gradient = 2 * np.average (X * (predict(X, w) - Y))
    return w_gradient

def train(X, Y, iterations, lr=0.00001):
    w = 0 
    for i in range(iterations):
        l = loss (X, Y, w)
        w_gradient = gradient(X, Y, w)
        w = w - w_gradient *lr
        # Update the height coefficient by multiplying the gradient and learning rate (lr).
    return w

X = np.array([ 161, 163, 178, 157, 188, 161])
Y = np.array([47.81, 54.92, 53.99, 45.84, 51.92, 58.01])

print("iterations,\tcoeff")
for its in range(5,155,5):
    height_coeff = train(X, Y, iterations=its)
    print('{},\t\t{}'.format(its, height_coeff))