cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
  • Register to attend Discovery Summit 2025 Online: Early Users Edition, Sept. 24-25.
  • New JMP features coming to desktops everywhere this September. Sign up to learn more at jmp.com/launch.
Choose Language Hide Translation Bar

How do I save the output report statistics of Neural predictive model performance? because they change every time I open the neural report.

I know that I can save the formula to predict the responses but to evaluate the model I need to keep the model statistics, and it changes every time I open the neural report. If someone can provide me with some guidance, I need to be able to open the neural report more than just once. 

1 REPLY 1
Victor_G
Super User

Re: How do I save the output report statistics of Neural predictive model performance? because they change every time I open the neural report.

Hi @Ayman_Aljazzar,

Welcome in the Community !

When you save the script of an analysis, you save the method and inputs to run the analysis. For "deterministic" analysis methods like linear regression, there are no randomness involved in the method, so you can expect to have the same results when re-running the analysis.
In the case of Neural Network (and most Machine Learning algorithms also involve randomness in the process), the weights of each neuron are randomly generated before being adjusted and optimized.
When you launch the Neural Network platform, you can specify a random seed : this will enable to start the neurons weights with the same values each time, so that your final results will be the same each time you run the scripts.

For more information about randomness in Machine Learning, you can read https://mindfulmodeler.substack.com/p/no-learning-without-randomness


Another option to save these metrics once you have run the Neural Network analysis is to right-click on the statistical metrics and use the option to save as a data table. You can then use Graph Builder or other platforms with this data table to visualize your results.

Hope this answer will help you,

Victor GUILLER

"It is not unusual for a well-designed experiment to analyze itself" (Box, Hunter and Hunter)

Recommended Articles