Overview of MARS® Regression

Use Multivariate Adaptive Regression Splines (MARS®) to create accurate predictive models for a continuous response with many continuous and categorical predictor variables.

The powerful basis functions that form models from MARS® Regression are highly adaptable so that the models capture significant departures from the linearity restraints of conventional multiple regression. MARS® Regression easily handles the complex data structure that often hides in high-dimensional data. In doing so, this approach to regression modeling effectively uncovers important data patterns and relationships that are difficult, if not impossible, for other regression methods to reveal.

Unlike predictive analytics models that use trees to form models, models from MARS® Regression have representations comparable to equations from conventional multiple regression. The relationships between the response variable and the individual predictors are easeier to understand with these equations.

MARS® Regression provides insights for a wide range of applications, including manufacturing quality control, drug discovery, fraud detection, credit scoring, and churn prediction. Use the results to identify important variables, to examine the relative effects of variables, and to predict response values for new observations. For example, real estate appraisers want to see how the sales price of urban apartments is associated with several predictor variables including the square footage, the number of available units, the age of the building, and the distance from the city center.

For descriptions of MARS® Regression and other predictive analytics models, go to Types of predictive analytics models in Minitab Statistical Software.

Where to find this analysis

To perform MARS® Regression, choose Predictive Analytics Module > MARS® Regression.
Note

This command is available with the Predictive Analytics Module. Click here for more information about how to activate the module.

When to use an alternate analysis

If you want to try a parametric model with a continuous response variable, use Fit Regression Model.

To compare the performance of a Random Forests® Regression model, use Random Forests® Regression.

To compare the performance of a TreeNet® Regression Model, use TreeNet® Regression

To compare the performance of multiple models simultaneously and produce results for the model with the best fit, use Discover Best Model (Continuous Response).