LASSO for variable selection

Dear MATLAB Learners,

I am working on variable selection for process monitoring.

I have only number of process response variables for n observations. My objective is to identify which response variables are more useful with the help of variable selection method.

How can I code this? I am attaching the following attachment for your reference.

Answer 1

[Sahas]

Hi @Nainsi Gupta,

I have provided you with some MATLAB Documentation links below which will help you through the implementation of LASSO for variable selection in MATLAB R2024a.

The following MATLAB Documentation below gives the basic understanding and ways of implementing “LASSO” function in MATLAB. Kindly go through the various examples of applying the “LASSO” function with distinct “parameters” and choose the best example as per your use case.

Lasso or elastic net regularization for linear models - MATLAB lasso - MathWorks India

The following MATLAB Documentation is an example to see ‘“LASSO” regularization technique in MATLAB. This example shows the working of ‘“LASSO” function and how it discards unnecessary predictors as well as a snippet for “lassoPlot” function for visualizing the regressed data.

Lasso Regularization - MATLAB & Simulink - MathWorks India

The below documentation link lists the various techniques of “Feature Selection” in MATLAB. Kindly go through the “Embedded Type Feature Selection” table for various alternatives to ‘“LASSO”.

Introduction to Feature Selection - MATLAB & Simulink - MathWorks India

I hope this helps you get started!

[Nainsi Gupta]

Dear Sahas,

These documentation of lasso requires X (predictor) as well as y (response) variable. But I don't have both datasets.