Is additive logistic regression equivalent to boosted decision stumps?

Question

Are additive logistic regression and boosted decision stumps (where a decision stump is a one-node decision tree) equivalent in some sense? I thought not, but if I google for "LogitBoost" algorithms, I find:

• http://weka.sourceforge.net/doc/weka/classifiers/meta/LogitBoost.html, which says it's a "class for performing additive logistic regression"
• http://rss.acs.unt.edu/Rdoc/library/caTools/html/LogitBoost.html, which says "Train logitboost classification algorithm using decision stumps (one node decision trees) as weak learners."

Answer 1

Boosted decision stumps is just a special case of generalized additive models (i.e. if the logistic loss function is used then, technically, one could call boosted decision stumps an additive logistic model). Having said that, people typically use specialized names for boosted models - for example Gradient Boosting Machine which also belongs to the class of generalized additive models (with decision tree as the base learner) and supports multiple loss functions including the logistic one.