If a primary interest is to conduct automatic variable selection, only LASSO can do that. A general guideline is that if it is thought that only a few variables are strongly related to the response, LASSO would be the better choice, since it would eliminate the unimportant features. On the other hand, if many predictors are thought to have at least a small degree of explanatory power, Ridge might be a better choice since it would allow all features to retain some influence in the model. However, it is usually advantageous to train both models and compare their performance, as depending on the dataset, one might perform better than the other.
When to use Ridge Regression vs Lasso?
Help us improve this post by suggesting in comments below:
– modifications to the text, and infographics
– video resources that offer clear explanations for this question
– code snippets and case studies relevant to this concept
– online blogs, and research publications that are a “must read” on this topic
Partner Ad
