Enter your data for Response Optimizer

Complete the following steps to perform response optimization. The steps depend on the type of model.

Predictive Analytics Module

Complete the following steps to specify the response to optimize.

  1. Go to the results for a model, then select Response Optimizer.
  2. If you have a categorical response, in Optimizer classes, select the classes to use for the goal. For a multinomial response, select one or more classes.
    For example, suppose that your goal is to maximize desirable responses. The response variable has 5 classes with 2 desirable responses: "Satisfied" and "Very Satisfied." Select "Satisfied" and "Very Satisfied" and select the Maximize goal in the next step.
  3. In the table, in Goal, select one of the following options for the response.
    • Do not optimize: Do not include the response in the optimization process.
    • Minimize: Lower values of the response are preferable.
    • Target: The response is optimal when values meet a specific target value.
    • Maximize: Higher values of the response are preferable.
    For more information, go to Which response optimization goal should I use?.
  4. In Target, for each response that has a goal of target, enter a target value. With a binary response, the target value must be between 0 and 1.

Stat menu except mixture designs

Complete the following steps to specify the responses to optimize.

Note

Only response variables with up-to-date models from the same type of analysis are in the list. If you do not see a response that you want, re-fit the model. For more information, go to Stored model overview.

  1. To perform this analysis in Minitab, go to the menu that you used to fit the model, then choose Response Optimizer. For example, if you fit a Poisson model, choose Stat > Regression > Poisson Regression > Response Optimizer.
  2. In the table, in Goal, select one of the following options for each response.
    • Do not optimize: Do not include the response in the optimization process.
    • Minimize: Lower values of the response are preferable.
    • Target: The response is optimal when values meet a specific target value.
    • Maximize: Higher values of the response are preferable.
    For more information, go to Which response optimization goal should I use?.
  3. In Target, for each response that has a goal of target, enter a target value. With a binary response, the target value must be between 0 and 1.

Mixture designs

Complete the following steps to specify the responses to optimize.

  1. To perform response optimization for a mixture design, choose Stat > DOE > Mixture > Response Optimizer.
  2. Move the variables that you want to include in the response optimization from the Available list to the Selected list.
  3. Under Model Fitted in, choose whether you want to refit the model in Proportions or Pseudocomponents. For more information, go to Amounts, proportions, and pseudo-components scales for representing the data in a mixtures design and What is a pseudo-component?.
  4. You must select Setup and specify the goal, boundaries, weight, and importance for each response variable before you can perform the analysis.