Scribd Logo
Upload

Welcome to Scribd!

  • 1 views

    Gradient Descent Optimization

    Uploaded by

    amoghakantak

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Gradient Descent Optimization

    Uploaded by

    amoghakantak
    1 views1 page

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    1 views1 page

    Gradient Descent Optimization

    Uploaded by

    amoghakantak

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    You are on page 1of 1

    Lab Session #8

    Gradient Descent Optimization using python

    Aim: To implement gradient descent optimization algorithms in python.

    Problem Definition: Download the dataset from the following link:


    https://www.kaggle.com/sazid28/advertising.csv . Use the dataset to implement gradient
    descent algorithms and project the costs for varying values of alpha.

    Theory: Gradient Descent is an optimisation algorithm which helps you find the optimal
    weights for your model. It does it by trying various weights and finding the weights which
    fit the models best i.e. minimizes the cost function. Cost function can be defined as the
    difference between the actual output and the predicted output. Hence, the smaller the cost
    function is, the closer the predicted output from your model is to the actual output. Cost
    function can be mathematically defined as:
    𝑦=𝛽+θnXn,
    where x is the parameters(can go from 1 to n), 𝛽 is the bias and θ is the weight
    While on the other hand, the learning rate of the gradient descent is represented as α.
    Learning rate is the size of steps taken by each gradient. While a large learning rate can give
    us poorly optimized values for 𝛽 and θ, the learning rate can also be too small which takes a
    substantial increment in number of iterations required to get the the convergence point(the
    optimal value point for 𝛽 and θ). This algorithm, gives us the value of α, 𝛽 and θ as output.

    You might also like