Multivariable Linear Regression with Gradient Descent Algo - Concept and Implementation from scratch

Опубликовано: 18 Декабрь 2021
на канале: Programming Epitome
2,700
50

Implementation of Linear Regression using Gradient descent algorithm with two feature variables from scratch. Detailed concept of Multi variable linear Regression, gradient descent algorithm, the loss function, mean squared loss and how it works, updating weights, Learning rate, derivatives of loss function to find slope to reach minima. Vectorized code using numpy to avoid multiple for loops.

Video Sections:

0:00 - Intro
02:36 Concept Linear Regression
08:19 Concept Gradient Descent, Mean squared Loss, Derivative, weights update, learning rate, pseudocode
20:46 Implementation of Linear Regression
22:30 Standardization of variables
23:06 Concatenation for Vectorized Code
27:38 Gradient Descent Code
31:39 Working of numpy.dot function
35:51 Gradient Descent Code
44:11 Plotting cost function
45:12 Animation Generation

Acknowledgements:
Dataset used can be download from here: https://drive.google.com/file/d/1NBoR...
Animation code was adapted from here: https://www.kaggle.com/ronaldtroncoso...
Other code help was taken from here: http://aimotion.blogspot.com/2011/10/...