Thanks for the response Claude!

One of the concerns I had when building out the pipeline is that there is a lot of output to the console. So when considering tracking each of the optimal models, a better logging system would be needed.

Moreover, the pipeline is setup to iterate through each configuration, therefore the final predictive model will show up multiple times in the loop. Tracking the best result for SVM vs. Random Forest vs. Logistic Regression is more complicated code that could clutter a post.

And yes the random seed should be set for consistent results.

Thanks again for your feedback!