Linear_Regression_0.py

import numpy as np
import matplotlib.pyplot as plt

def fit_transform_poly(x,w,n,b):
    '''Evaluates f=W*x^n+b'''
    p_Matrix=[]
    for weights,power in zip(w,n):
        p_Matrix.append([weights*x**power for x in x])
    return [sum(i)+b for i in zip(*p_Matrix)]

coef=[1,-.1] # Define the Poly. Weights.
power=[1,2] # Define the order of the Polynomial 
bias=0 # Define the bias
n_samples=100
x=np.linspace(0,5,n_samples) # Create linear space of 100 points between (0,5).
std=0.3 # Standard deviation of the noise.
e=np.random.normal(0,std,len(x)) # Generate noise.

assert len(coef)==len(power), " The number of coef. (W)\
 must equal the power (n)"
 
f=fit_transform_poly(x,coef,power,bias) + e #Eval. the poynomial and add noise.