Multiple Regression
Multiple Regression |
|
|
How To... > Work with Statistical Objects > Build a Linear Model > Multiple Regression Most things you want to explain or predict aren’t simple; they depend on a multiplicity of factors. Statisticians have developed techniques for modeling under such circumstances, and one of these techniques is multiple linear regression. The basic idea is that there is one continuous attribute, the response, whose values are hypothesized to depend linearly on one or more other numeric attributes, called predictors. The goal of building a multiple linear regression model is to come up with an equation that does a good job of predicting the response from the predictors. If R is the response attribute and P1, P2, …, PN are predictors, this equation is of the form:
R = A0 + A1*P1 + A2*P2 + … + AN*PN, where A0 through AN are coefficients determined during the model-building process. The basic steps are to create a linear model object, set it to Multiple Regression, then drop attributes on the prompts. See this topic’s Subsections from more detail. Subsections Adding, Removing, and Moving Predictor Attributes: The Goal of Model Building Changing Multiple Regression Output |
