Is the combination of labeled and unlabeled data the way to deploy prediction models?

virpi_hotti

Is the combination of labeled and unlabeled data onlt way to deploy prediction models generated by auto model in RapidMiner Studio?

sgenzer

hello @virpi_hotti welcome to the community. I think I just talked to a colleague of yours today on chat? So what you describe IS basically the definition of predictive analytics. That's the way it's done. Maybe you can explain a bit more what you're doing? Or maybe I did not not understand your question?

Scott

virpi_hotti

For example, there are two samples Titanic Training and Titanic Unlabeled - I combined them in Excel and then I import one file and generate parsimonious models for predictions. If you have a better ideas to deploy and compare the parsimonious models - please, let me know.

IngoRM

Hi,

No, but it is probably the easiest way to deploy a model (although not the best or most time-efficient).  There are probably a dozen or so ways to deploy models in RapidMiner.  I recommend a visit in our Academy at http://academy.rapidminer.com to learn more about them.

Here are a couple of pointers:

• https://academy.rapidminer.com/learn/video/scoring-intro
• https://academy.rapidminer.com/learn/video/applying-the-model
• https://academy.rapidminer.com/learn/video/scoring-demo

There is in fact a whole part of the catalog dealing with model deployment: https://academy.rapidminer.com/catalog?labels=["Phase"]&values=["Deployment"]

Hope this helps,

Ingo