gradient descent for linear regression to find formula for linear function

task1.py

# todo: find functions formulas using gradient descent
# todo: check if you have
# t1 answers
# h1 = 1 2
# h2 = 10 20

import numpy as np

features = np.array([
    [1, 1, 1, 1],
    [-1, 2, 3, 10]
])
Y = np.array([-1, 5, 7, 21])
T = np.array([1, 10])


def costDerivative(T, features, Y):
    return T.dot(features) - Y


def costDerivativesPerFeature(T, features, Y):
    return map(
        lambda feature: costDerivative(T, features, Y).dot(feature),
        features
    )


def gradientDescent(nextT, features, Y):
    samplesLength = features.shape[1]
    step = 0.05
    prevT = np.array([0, 0])
    count = 0

    while any(prevT != nextT):
        count += 1
        prevT = nextT
        nextT = np.array(list(map(
            lambda t, costDerivativeForFeature: t - 1 / samplesLength * step * costDerivativeForFeature,
            nextT, costDerivativesPerFeature(nextT, features, Y)
        )))

    return nextT, count


print(gradientDescent(T, features, Y))