# Regularization

## What is Regularization :

Regularization is a technique used in machine learning to prevent overfitting, which is when a model fits too closely to the training data and performs poorly on unseen data. There are two main types of regularization: L1 and L2.
L1 regularization, also known as Lasso, is a method that adds a penalty term to the objective function, which is the function being optimized in the model. The penalty term is the absolute value of the weights multiplied by a hyperparameter, which is a parameter that must be set by the user. This hyperparameter determines how much the weights are penalized. L1 regularization results in a sparse model, meaning that some weights are set to zero, which can be beneficial for feature selection.
For example, consider a linear regression model that is used to predict the price of a house based on its size and number of bedrooms. Without regularization, the model might fit the training data very well, but it might not generalize to unseen data. L1 regularization can be used to prevent overfitting by adding a penalty term to the objective function that is proportional to the absolute value of the weights. This would result in some weights being set to zero, which would reduce the complexity of the model and make it more robust to unseen data.
L2 regularization, also known as Ridge, is another method that adds a penalty term to the objective function. The penalty term is the square of the weights multiplied by a hyperparameter, which is similar to L1 regularization. However, unlike L1 regularization, L2 regularization does not result in sparse models. Instead, it shrinks the weights towards zero, but does not set them to zero. This can be beneficial for preventing overfitting, but it can also reduce the model’s ability to capture important relationships in the data.
For example, consider a neural network that is used to classify images of dogs and cats. Without regularization, the model might fit the training data very well, but it might not generalize to unseen data. L2 regularization can be used to prevent overfitting by adding a penalty term to the objective function that is proportional to the square of the weights. This would shrink the weights towards zero, but it would not set them to zero. This would reduce the complexity of the model and make it more robust to unseen data.
In conclusion, regularization is a technique used in machine learning to prevent overfitting and improve the generalization of a model. There are two main types of regularization: L1 and L2. L1 regularization adds a penalty term to the objective function that is proportional to the absolute value of the weights and results in a sparse model. L2 regularization adds a penalty term to the objective function that is proportional to the square of the weights and shrinks the weights towards zero, but does not set them to zero. Regularization can be beneficial for improving the performance of a model on unseen data, but it can also reduce the model’s ability to capture important relationships in the data.