LASSO stands for Least Absolute Selection Shrinkage Operator, and shrinkage is specified as a parameter constraint. The purpose of lasso regression is to find the collection of predictors with the lowest prediction error for a quantitative response variable. The approach works by placing a restriction on the model parameters, causing the regression coefficients for some variables to decrease toward zero.
Variables having a regression coefficient of 0 after shrinking are removed from the model. Variables with non-zero regression coefficients have the strongest relationship with the response variable. Explanatory variables might be quantitative, categorical, or a combination of the two. This lasso regression analysis is essentially a shrinkage and variable selection strategy that assists analysts in determining which predictors are most relevant.
THE MAIN CODE OF LASSO REGRESSION
Full Code Exercise of Lasso Regression
In this part, we will show how to apply the Lasso Regression technique. First, let's look at a typical regression dataset. We'll be using the data from technical spec of cars. The dataset is downloaded from UCI Machine Learning Repositor.
The dataset can be downloaded here.
from sklearn.linear_model import Lasso