Multifactor dimensionality reduction

<p>
Multifactor dimensionality reduction (MDR) is a statistical approach, also used in <a href="/facts/Machine_learning/e0w0XJTu">machine learning</a> automatic approaches, for detecting and characterizing combinations of <a href="/facts/Attribute_(computing)/EpQSmn02">attributes</a> or <a href="/facts/Independent_variable/dINfzIF2">independent variables</a> that interact to influence a dependent or class variable. MDR was designed specifically to identify nonadditive <a href="/facts/Interaction_(statistics)/Kp0Cab6w">interactions</a> among <a href="/facts/Discrete_random_variable/TwTBXnLT">discrete</a> variables that influence a <a href="/facts/Binary_numeral_system/a6JDSRPs">binary</a> outcome and is considered a <a href="/facts/Nonparametric/svTDJHTb">nonparametric</a> and model-free alternative to traditional statistical methods such as <a href="/facts/Logistic_regression/mGj6mc8y">logistic regression</a>.
</p><p>The basis of the MDR method is a constructive induction or <a href="/facts/Feature_engineering/VkNJFAFb">feature engineering</a> algorithm that converts two or more variables or attributes to a single attribute. This process of constructing a new attribute changes the representation space of the data. The end goal is to create or discover a representation that facilitates the detection of <a href="/facts/Nonlinear_regression/8mMuJSe8">nonlinear</a> or nonadditive interactions among the attributes such that prediction of the class variable is improved over that of the original representation of the data.
</p>

Multifactor dimensionality reduction open-in-new

Multifactor dimensionality reduction