Cornerstone’s Multiple Regression analysis module allows you to create and refine models to understand your data.

Explore Data Relationships

In today’s competitive business environment, having the ability to understand the relationships among data can be key to your success.

In manufacturing for example, determining how yields vary as the equipment varies can significantly impact your bottom line.

Regression analysis is a powerful technique for finding the relationships between one or more predictor variables and a response variable.

A Powerful Tool to Understand Your Data

Cornerstone’s Multiple Regression module allows you to model a response (process output) as a function of the level of one or more predictors (process inputs). You can then use the model to obtain information that can help you improve the process you are studying.

Cornerstone’s Multiple Regression analysis module is designed to be easy to use and learn. The module provides a user-friendly environment for building, refining, and using models to help you understand your data and make informed decisions.

Dynamic graphs and an intuitive user interface provide support for:
· Developing linear, quadratic and cubic models
· Adding and removing variables and terms
· Performing residual analysis
· Making predictions

Extend Your Application

Multiple Regression’s functionality is part of a suite of data analysis modules. The functionality in all of Cornerstone’s analysis modules is accessible through the Cornerstone™ Extension Language (CEL). CEL allows you to extend Cornerstone and to create custom, site-specific applications.

Features

Model Creation and Refinement

· Create multiple regression model
· Add or remove variables from the model
· Add or remove higher order terms
· Select outliers manually or statistically
· Change set of observations used in model
· Perform automatic stepwise regression

Model Evaluation

· Plot response vs specified predictor
· Table of estimated regression coefficients
· Table of model coefficient correlations
· Goodness-of-fit statistics
· Model anova table
· Component anova table
· Residual analysis
· Histogram of residuals distribution
· Normal probability plot of residuals
· Residuals vs. actual values of any predictor
· Residuals vs. fitted response values
· Table of residuals and fitted values for each observation

Model Use and Visualization

· Plot response vs. one predictor as scatter or box plot
· Plot response vs. two predictors as surface or contour plot
· Table of predicted values of the response
· Specify and show confidence levels

Guidance and Interpretation