Make better predictions with powerful regression procedures!

When you need to build better predictive models than you get using simple linear regression, you need SPSS Regression Models. You can apply more sophisticated models with SPSS Regression Models' wide range of nonlinear modeling procedures. And, you can use SPSS Regression Models to predict behaviors, actions or attitudes with multidimensional scaling.

The new multinomial logistic regression procedure predicts a categorical outcome such as "primary reason for web use". The categories are: a) work only, b) shopping only, c) both working and shopping, and d) neither (reference category). We can see that search engine use was a better predictor of "shopping only" than print media use.

Use SPSS Regression Models for:

• Market research
  Studying consumer buying habits
• Medical research
  Studying response to dosages
• Loan assessment
  Analyzing good and bad credit risks
• Institutional research
  Measuring academic achievement tests
• And much more

Free yourself from data constraints such as yes & no answers
Use the new Multinomial Logistic Regression procedure to predict categorical outcomes with more than two categories. For example, what predicts whether the customer buys product A, product B or product C. Your predictors can be ither categorical or continuous level.

Classify your data into two groups easily
Use Binary Logistic Regression to predict dichotomous variables such as buy or not buy, vote or not vote. This procedure offers many stepwise methods to select the important continuous or categorical covariates which best predict your response variable.

Control your model
Get more control over your model and your model expression with the Constrained and Unconstrained Nonlinear Regression procedures. These procedures provide two methods for estimating parameters of non-linear models. The Levenberg-Marquardt algorithm analyzes unconstrained models. The sequential quadratic programming algorithm lets you specify constraints on parameter estimates, provide your own loss function, and get bootstrap estimates of standard errors.

Loosen your assumptions
When your data do not meet the statistical requirements for ordinary least squares, use Weighted Least Squares (WLS) and Two-Stage Least Squares (2SLS). Give more weight to measurements within a series with WLS. 2SLS helps control for correlations between predictor variables and error terms that often occur with time-based data.

Find the best stimuli
Perform Probit and Logit response modeling to analyze the potency of responses to stimuli such as medicine doses, prices or incentives. Probit evaluates the value of the stimuli using a Logit or Probit transformation of the proportion responding.

Sales of SPSS Regression Model is restricted to the U.S. and Canada only