I did not use it for a learning project with real data, I used it maybe 3 years ago to learn more about data mining/science and as an alternative to python/sklearn which I had previous experience with. I only did tutorials with supplied data, and the goal was to try the platform.
As I wrote in another thread before using any ML program think of multiple comparisons and read this classic article from PAL blog.
Can you input external strategy code in RapidMiner to configure best meta-strategy? I do this with custom code someone wrote for me for strategies I get from DLPAL S usually about 50 to 100 for each contract I trade. As it turns out meta-strategies developed by genetic optimization are superior when compared to naive ensembles.
If I understand you correctly, I believe so. https://docs.rapidminer.com/latest/developers/creating-your-own-extension/
lol RapidMiner, that is such a blast from the past. I remember trying it when I hadn't a clue about machine learning. You can guess how that went. I just don't see what the point of this software is. Once you know enough about machine learning to use it, you will just end up using a programming language.