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

file_path = "slr12.xls"

def h_theta(theta,x):
    return theta[0]+theta[1]*x

def f_costo(dataset,theta):
    size=dataset.shape[0]
    return np.mean((np.sum( np.power(h_theta(theta,x)-y,2) for x,y in dataset.get_values()))/2)

def minimize(dataset,theta):
    size=dataset.shape[0]
    a=0.001 #learning rate
    tmp=[0,0]
    while f_costo(dataset,theta)>10:
            tmp[0]=theta[0]-a*np.mean(np.sum(h_theta(theta,x)-y for x,y in dataset.get_values()))
            tmp[1]=theta[1]-a*np.mean(np.sum((h_theta(theta,x)-y)*(x) for x,y in dataset.get_values()))
            theta[0]=tmp[0]
            theta[1]=tmp[1]
            print(theta)
    print(f_costo(dataset,theta))
    return theta

dataset = pd.read_csv(file_path)
theta = [1,1]
print(f_costo(dataset,theta))
theta = minimize(dataset,theta)
print(theta)

plt.figure()
plt.scatter(dataset.X,dataset.Y)

dataset