Menu
Home
People
Places
Arts
History
Plants & Animals
Science
Life & Culture
Technology
Reference.org
Additive model
open-in-new
<p>In <a href="/facts/Statistics/DNPsKGYU">statistics</a>, an additive model (AM) is a <a href="/facts/Nonparametric_regression/3MuUSUFV">nonparametric regression</a> method. It was suggested by <a href="/facts/Jerome_H._Friedman/brTdEkVe">Jerome H. Friedman</a> and Werner Stuetzle (1981) and is an essential part of the <a href="/facts/Alternating_conditional_expectations/c398UNL3">ACE</a> algorithm. The <i>AM</i> uses a one-dimensional <a href="/facts/Smoothing/2svFjsbI">smoother</a> to build a restricted class of nonparametric regression models. Because of this, it is less affected by the <a href="/facts/Curse_of_dimensionality/dk7SnY6L">curse of dimensionality</a> than a <i>p</i>-dimensional smoother. Furthermore, the <i>AM</i> is more flexible than a <a href="/facts/Linear_regression/5n998IhK">standard linear model</a>, while being more interpretable than a general regression surface at the cost of approximation errors. Problems with <i>AM</i>, like many other machine-learning methods, include <a href="/facts/Model_selection/R11QK2GC">model selection</a>, <a href="/facts/Overfitting/5xnFLcMg">overfitting</a>, and <a href="/facts/Multicollinearity/bQXzB6sA">multicollinearity</a>. </p>