How to get response optimizer with DOE analysis and neural modeling? Thanks

Question

At Polytechnic Montréal (engineering faculty) we have statistica licence. I can perform DOE and Neural analysis but still not able to get response optimizer (such as with Minitab) to get independant variable settings for a proposed dependant variable. Thanks for your support. Jean-Paul Lemarquis

Answer 1

To address your question about using STATISTICA for response optimization with DOE analysis and neural modeling, you may need to follow these steps:

1. DOE Analysis: Ensure that you have performed a thorough Design of Experiments analysis within STATISTICA. This typically involves creating an experimental design, conducting the experiments, and entering the response data into the software.

2. Neural Network Modeling: After DOE, develop a neural network model using the data you've collected. This step involves training the neural network to understand the relationship between your independent variables (factors) and the dependent variable (response).

3. Optimization:

   • Unfortunately, STATISTICA may not have a direct "Response Optimizer" feature equivalent to Minitab. However, you can use its "What-If Analysis," "Simulated Annealing," or "Genetic Algorithms" within the neural networks toolbox or other model optimizing modules to achieve a similar outcome.
   • You may need to manually adjust the independent variables to see how changes influence the response, using simulation and the neural model's prediction capabilities to guide you.

4. Seek Additional Resources: If you are not able to find a direct response optimizer, consider reaching out to Statistica technical support or leveraging online communities and forums where users of the software may have shared workarounds or scripts.

Make sure you have the correct licenses and modules active in STATISTICA to access all necessary features. If you require more specific guidance, consulting the software's documentation or support services will be helpful.