Statistical Learning from a Regression Perspective

John Maindonald
2009 Journal of Statistical Software  
Statistical learning is a name for methods that experiment with many different models for the data, and which have been a strong focus of data mining and machine learning traditions of data analysis. Here, regression encompasses classification as well as regression with a continuous or ordinal outcome variable. Regression splines and regression smoothers are the point of departure, before moving on to other methodologies -classification and regression trees, bagging, random forests, boosting
more » ... orests, boosting and support vector machines. Berk suggests that statistical learning may be the "muscle car" version of exploratory data analysis. Presumably the "muscle" is computing muscle. I would prefer to say that statistical learning is a "muscle car" version of regression methodology that emphasizes exploratory uses.
doi:10.18637/jss.v029.b12 fatcat:3n7q6idq35egppfzatxaj5nv6m