gradient descent algorithm in machine learning

Question

I am beginner in machine learning.ihave problem in gradient descent algo.in the code, mentioned below, my doubt is during

first iteration value of x will be 1

second iteration value of x will be 2

third iteration value of x will be 3

fourth iteration value of x will be 4

fifth iteration value of x will be 5

then what will be the value of x for iterations 6 to 9999???

import numpy as np

import matplotlib.pyplot as plt

%matplotlib inline

def gradient_descent(x,y):

    m_curr = b_curr = 0
    rate = 0.01
    n = len(x)
    plt.scatter(x,y,color='red',marker='+',linewidth='5')
    for i in range(10000):
        y_predicted = m_curr * x + b_curr
        plt.plot(x,y_predicted,color='green')
        md = -(2/n)*sum(x*(y-y_predicted))
        yd = -(2/n)*sum(y-y_predicted)
        m_curr = m_curr - rate * md
        b_curr = b_curr - rate * yd

x = np.array([1,2,3,4,5])

y = np.array([5,7,9,11,13])

gradient_descent(x,y)

Please format your code block properly. – desertnaut Jan 10 '19 at 17:16 — desertnaut, Jan 10 '19 at 17:16
You are not modifying `x`??? So it will stay the same. – Matthieu Brucher Jan 10 '19 at 18:15 — Matthieu Brucher, Jan 10 '19 at 18:15

score 0 · Answer 1 · answered Jan 10 '19 at 18:39

0

Three variations for calculating batch gradient

Batch
Stochastic
Mini-batch

I believe you are trying to implement the Batch variation. If that is the case you need to do the following

change the for loop to run until the size of the x or y. i.e. len(x)
call gradient_descent(x,y) in another for-loop with 10000 (or how many number of iterations you want to run). You should probably use a error function to determine this

answered Jan 10 '19 at 18:39

Aravind Yarram

78,777
46
231
327

Aravind R. Yarram ,in case of stochastic , len(x) should be equal to number of iteration??? – Jaspreet Kaur Bains Jan 11 '19 at 03:19

gradient descent algorithm in machine learning

1 Answers1