What is it about?
In high dimensional data where the number of variables is enormous and even much larger than observations, the multicollinearity phenomenon is inexorable and can be handled by penalized regression methods such as Ridge, Lasso, Elastic Net, and SCAD.
Featured Image
Photo by Stephen Phillips - Hostreviews.co.uk on Unsplash
Why is it important?
Our empirical study concludes that, in all correlation values, the performance prediction of the elastic net model outperforms the lasso model. Ridge fails to produce a better prediction model compared to the two others. In all levels of correlation, the variance of error resulting from all methods is quite similar.
Perspectives
I hope this article can help people in handling high dimensional data where ordinary least squared failed to solve
Deiby Tineke Salaki
Universitas Sam Ratulangi
Read the Original
This page is a summary of: Performance of penalized regression in predicting high dimensional data with various correlation values, January 2023, American Institute of Physics,
DOI: 10.1063/5.0118406.
You can read the full text:
Contributors
The following have contributed to this page
