[MXML-3-02] Linear Regression [2/7] - Overfitting and Regularization, LASSO and Ridge

In this video, we will cover the Overfitting and Regularization.

In the last video, we looked at the basics of Ordinary Least Squares. In this video, we will look at the LASSO and Ridge regularization methods to prevent overfitting. And we will implement the Linear Regression using the optimization class from scipy library.

Overfitting occurs when the regression line fits the training data so well that it does not fit the test data, which is new or unexperienced data. The regularized regression line slightly reduces performance on the training data but improves performance on new data. On the other hand, underfitting occurs when the regression line is too simple to describe both the training and test data well. Regularization is a key technique that prevents overfitting by adding a penalty term to the model's loss function.

#LinearRegression #Regularization #LASSO #Ridge #Overfitting #Underfitting

[KushJuvekar-jf]

Is w^2 used to minimize a convex function as opposed to w^n where n is any natural number and does not guarantee a convex function?