live.py

# -*- coding: utf-8 -*-
"""
Created on Mon Oct 23 08:08:21 2017

@author: Arafat
"""
"""We are basically dealing with gradient descent"""

import numpy as np

def compute_error_for_given_points(b, m, points):
    totalError = 0
    for i in range(len(points)):
        x = points[i,0]
        y = points[i,1]
        totalError += (y - (m*x + b))**2
    return totalError / float(len(points))

def step_gradient(b_current, m_current, points, learningRate):
    b_gradient = 0
    m_gradient = 0
    N = float(len(points))
    for i in range(len(points)):
        x = points[i, 0]
        y = points[i, 1]

        b_gradient += -(2/N) * (y - (m_current*x + b_current))
        m_gradient += -(2/N) * x * (x - (m_current*x + b_current))
        
        #same thing
        """b_gradient += -(2/N) * y
        m_gradient += -(2/N) * x * x"""
        
    new_b = b_current - (learningRate * b_gradient)
    new_m = m_current - (learningRate * m_gradient)
    return [new_b, new_m]

def gradient_descent_runner(points, starting_b, starting_m, learning_rate, num_iterations):
    b, m = step_gradient(starting_b, starting_m, np.array(points), learning_rate)
    return [b, m]

def run():
    points = np.genfromtxt('data.csv', delimiter=',')
    
    #hyper-parameters
    learning_rate = 0.0001
    
    #y = mx + b (slope formula)
    initial_b = 0
    initial_m = 0
    
    num_iterations = 1000
    [b,m] = gradient_descent_runner(points, initial_b, initial_m, learning_rate, num_iterations)
    print(b)
    print(m)

if __name__ == '__main__':
    run()