How to parameterise the training process

divya_das

Suppose I have a Process that takes an input training file & saves a Model. What if I want to re-use the same process for multiple customers. The input training file and model name will change depending on the customer. Can we parameterize the training process? One process for each customer will be difficult to manage in the long run, so we want to re-use the same process. 

kypexin

Hi @divya_das

Maybe I am not getting righ what do you mean by 'parameterizing the process', but to my understanding you run the process locally with a certain dataset and then save the resulting model, so why can't you just straightforwardly use the same process with different customers data? Once you have built a unified pipeline for data preparation and model training, you just can use it again and again with different datasets.

IngoRM

Hi,

Another thing for "parameterizing processes" is the concept of macros in RapidMiner.  Check out the following resources, they may help:

https://academy.rapidminer.com/learn/video/macros-introduction

https://academy.rapidminer.com/learn/video/macros-demo

https://community.rapidminer.com/discussion/32966/how-to-use-macros

https://docs.rapidminer.com/latest/studio/getting-started/macros.html

And finally the operator "Execute Process" may be helpful as well: https://docs.rapidminer.com/latest/studio/operators/utility/execute_process.html

Hope this helps,
Ingo

divya_das

Thank you for your help. I will try to create a re-usable process using macros and "execute process" operator.